OpenAI's Trillion-Dollar Gamble: Massive Investments in AI Computing Power Raise Bubble Concerns

OpenAI's Altman calls AI sector 'bubbly', but says we shouldn't worry - here's why

The AI boom has fueled an unbridled demand for computing power. The popularity of generative AI technologies has led to an unprecedented amount of investment in AI-related companies, infrastructure, and solutions. This has sparked concerns about the possibility of an AI bubble bursting and true ROI on investments, but OpenAI's CEO, Sam Altman, is now pushing back against that idea. At OpenAI DevDay 2025, Altman told the press in a Q&A that while the AI sector has some "bubbly" qualities, which may lead to overinvestment, the investment cycle is a normal part of any technological revolution, and it doesn't contradict the value that being built in AI space. Also: OpenAI DevDay event: Agent Kit, Apps SDK, ChatGPT, and more "People will overinvest in some places," said Altman. "There will be numerous bubbles and corrections over that period, but what I don't think [is that] this is totally divorced from reality -- there's a real thing happening here." These comments come at a time when the ROI on AI has yet to be proven, with multiple studies suggesting that despite hefty investments, the return remains unclear. For example, an MIT study found that 95% of enterprises attempting to harness the technology aren't seeing measurable results in revenue or growth. While Gaurav Gupta, Gartner's VP Analyst in Emerging Trends and Technologies, acknowledges that AI ROI has yet to be proven, he finds that there is a long road ahead until AI solutions, such as LLMs, reach their full potential, which is in part fueling further investments. "We have been talking about the AI bubble, where enterprises are finding it hard to achieve ROI beyond initial productivity gains with AI," said Gaurav Gupta, Gartner's VP Analyst in Emerging Trends and Technologies. "On the other hand, you can see hyperscalers, frontier labs, and advertising companies continue to spend to get access to more compute. This tells us that a lot more work still needs to be done on LLMs and a race towards AGI -- hence all the crazy demand for compute." Also: Despite AI-related job loss fears, tech hiring holds steady - and here are the most in-demand skills A prime example of a hyperscaler investment was made on Monday, ahead of DevDay, when AMD announced a new partnership with OpenAI. In this partnership, OpenAI agreed to utilize multiple generations of AMD Instinct GPUs to power its AI infrastructure, utilizing six gigawatts of power. The first one-gigawatt deployment of AMD Instinct MI450 GPUs is scheduled to take place in the second half of 2026. As part of the deal, AMD granted OpenAI a warrant for up to 160 million shares, equivalent to 10% of the outstanding shares of AMD common stock. This deal is only one of many OpenAI has made recently to meet GPU demand. In September, an Nvidia and OpenAI partnership allowed OpenAI to deploy 10 gigawatts of Nvidia systems, representing millions of GPUs. As part of the partnership, Nvidia agreed to invest up to $100 billion in OpenAI. Around the same time, OpenAI expanded its deal with GPU cloud provider CoreWeave to approximately $22.4 billion. Gupta added that the willingness of these massive companies to make significant investments in OpenAI supports the premise that OpenAI can deliver a lot of value in the AI space in the years ahead. "OpenAI is still a private company with massive valuation, but big companies like Nvidia and AMD continue to make deals with OpenAI -- it tells a lot about what they perceive about OpenAI -- promising future potential," said Gupta. The voracious industry-wide appetite for GPUs is fueled by a need for compute, which not only powers current offerings but also enables the development of new ones. Recent product releases have continued to highlight the need for more compute power. Also: AI lifts some software stocks, leaves others behind - who's winning and losing and why For example, Greg Brockman, OpenAI Vice President, said that the number one lesson from the Sora video generator release is the need for additional compute power. Another example is the new Pulse feature, which provides users with a personalized digest of news from across the web based on their chat activities with ChatGPT. Although it is a highly beneficial feature, its implementation is limited to Pro subscribers due to the computational demands of the feature. Ultimately, for every additional GPU investment, Altman said, there is a present ROI there. "We can still monetize every GPU we get our hands on super well; I think that will keep going for a long time," added Altman. "The degree to which we could have 10x of compute, we could build so many more products and offer so many more services people would love."

OpenAI: Here's Why Spending Billions on GPUs Makes Sense

When he's not battling bugs and robots in Helldivers 2, Michael is reporting on AI, satellites, cybersecurity, PCs, and tech policy. Spending hundreds of billions on AI might scream "tech bubble," but OpenAI insists its investments are all about keeping up with the relentless demand. "I think that we are still very much at the .01% of where we're going," OpenAI President Greg Brockman told journalists during a Q&A session at the DevDay conference on Monday. The company has recently made massive deals with Nvidia, Oracle, and now AMD so it can build energy-hungry gigawatt data centers to power next-generation AI. The spending may sound over the top -- especially since it'll mean building millions of new enterprise GPUs and even new power plants -- to keep such data centers running. But in OpenAI's view, the company is merely scratching the surface of its full potential. "The degree to which, if we had a 3x, 10x more compute, we would just build -- offer so many more products, offer more services that people would love to consume more," OpenAI CEO Sam Altman said during the Q&A. "This can go very far." One of the company's latest projects is the Sora app, an AI video generator that might one day rival TikTok. Still, other studies have cast doubt on whether generative AI can substantially improve worker productivity. Nevertheless, Brockman said the industry should embrace the notion that AI will become the main growth engine for the economy by boosting worker productivity. "I think the thing you have to believe for all of this to make sense is something that we've believed for a long time: Which is that AI is going to become, probably not in the too distant future, the fundamental driver of economic growth," he said. "If you believe that, then you start to see, asking how much compute you want is a little bit like asking how much workforce you want," he added. Brockman also went as far to say the AI industry could one day try to serve every user on the planet, or 8 billion people, which makes the company's investment seem small in comparison. "Our Nvidia deal, for example, is like for an order of 5 million GPUs," he said. "And then you start to realize there are almost 10 billion humans on the planet. We're talking about 1000x too few GPUs for our already quite large build." Still, Altman at one point said, "It's tempting to write the bubble story... In fact, there are many parts of AI that I think are bubbly," alluding to huge, but questionable, investments going to other AI startups. However, OpenAI's CEO said it's clear that the company's own technology has been leading to tangible benefits for users. On Monday, OpenAI revealed ChatGPT now has over 800 million weekly users -- up from 100 million back in 2023. The company is still navigating how to generate revenue from all of its users. However, OpenAI COO Brad Lightcap justified the wait-and-see approach. "If you had said that the business model of ChatGPT was going to be a consumer subscription for $20 a month, people would say that you're crazy. Because consumers don't pay for software," he said. "Yet now we got tens of millions of users that pay for it." "There has to be this exploration process for how these things work," Lightcap added. "There's going to be over investment, there's going to be incorrect investment. But there's going to be a lot of correct investment." Disclosure: Ziff Davis, PCMag's parent company, filed a lawsuit against OpenAI in April 2025, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.

What could burst the AI bubble?

University of Salford and University of Leeds provide funding as founding partners of The Conversation UK. Some of the world's biggest tech firms have soared in value over the last year. As AI evolves at pace, there are hopes that it will improve lives in ways that people could never have imagined a decade ago - in sectors as diverse as healthcare, employment and scientific discovery. OpenAI is now worth US$500 billion (£373 billion), compared with US$157 billion last October. Another firm, Anthropic, has almost trebled its valuation. But the Bank of England has now warned of a possible rapid "correction" due to its concerns about these staggering valuation rises. The question is whether these values are realistic - or based on hype, excitement and unfounded optimism for the potential of AI. Put simply, is AI's value today a product of what AI will do in future or what people hope it may do? Ultimately, we will only really know if it's a bubble if it bursts - though the warning signs are evident today. With hindsight, many things that happen in a bubble may sound exceedingly optimistic. If you take many headlines and replace the word AI with the word computers it often sounds a lot more naive. But, predicting the path of technological change is hard. Back in 2000 the Daily Mail declared the internet could be a passing fad. Just a few months earlier the dotcom boom had peaked. A burst bubble may not change the end of the journey. The internet was not a passing fad. However, bubbles are extremely disruptive and affect people in very real ways. Stocks fall, pensions suffer, unemployment rises and investment is wasted. Real potential is crowded out in the hype and mania to focus all investment in a small number of stocks and firms. Right now, we have the first sign of a bubble - a rapid rise in valuations. If these correct and fall we will have a bubble. If these valuations continue to rise we could be seeing a new sustained market that is focused on the technology of the future. Of course, it might be that these valuations plateau. What happens then depends on whether people have invested in the belief that prices will always rise. Consider a situation where people believe - as the Bank of England does - that AI firms' valuations may be "stretched". It's helpful to consider what these valuations are based on. Investment is simply a bet that AI increases profitability for the firms involved. These massive valuations are bets that AI will hugely increase future profitability. In some cases these are bets that AI will improve in capabilities towards some kind of "artificial superintelligence" that can do everything a human can do - or more. This could raise the living standards of everyone on Earth. Leading computer scientist Stuart Russell estimates the value of that at US$14 quadrillion - investors are buying a claim on that outcome too. If investors begin to fear that AI profits won't materialise then they will try to get their money back. This realisation can appear quite suddenly and can be prompted by seemingly minor events. It doesn't require a big needle to pop a bubble. A US article published in March 2000 warned that internet companies were fast running out of money. This caused many people to rethink their investments At this stage of the bubble, investment excitement had spread to everyday investors. These regular people balanced their fear of missing out with a fear that they were investing in something new that they didn't know much about. For many, an article in a popular magazine suggesting they may have made a mistake tipped the scales towards caution. They began to sell their dotcom stocks. In search of profit It may come as a surprise to some that, despite its increasing valuations, OpenAI does not yet make a profit. It may require ten times more revenue to do so. A US$500 billion valuation is quite something for a company that reportedly lost US$7.8 billion in the first half of this year. Some of this value appears to flow from a new deal between OpenAI and Nvidia where Nvidia will invest in OpenAI and OpenAI will buy Nvidia chips. This circular financing keeps everything afloat for now, but at some point investors will need to see returns. AI firms more generally do not appear to be profitable at the moment. Investors are not putting their money into today's losses - they are betting on an AI future. It is of course perfectly feasible that AI firms will develop business models to increase their profitability. OpenAI is exploring advertising options and allowing chatbots to recommend products. Using AI to deliver these messages is a viable option, though they will have to avoid the tricks and manipulations associated with online platforms, such as when hotel websites announce that rooms are about to sell out. We believe that AI can increase the power of these manipulations and we wonder how persuasive chatbots may be in their recommendations. However, the big four - Meta, Alphabet, Microsoft and Amazon - are this year spending the equivalent of the GDP of Portugal on AI infrastructure. This is not investment in new targeted ads, it is investment in an AI future. The bubble will burst if and when this future is in doubt.

OpenAI, Nividia Fuel $1 Trillion AI Market With Web of Circular Deals

Two weeks ago, Nvidia Corp. agreed to invest as much as $100 billion in OpenAI to help the leading AI startup fund a data-center buildout so massive it could power a major city. OpenAI in turn committed to filling those sites with millions of Nvidia chips. The arrangement was promptly criticized for its "circular" nature. This week, undeterred, OpenAI struck a similar deal. The ChatGPT maker on Monday inked a partnership with Nvidia rival Advanced Micro Devices Inc. to deploy tens of billions of dollars worth of its chips. As part of the tie-up, OpenAI is poised to become one of AMD's largest shareholders.

OpenAI's computing deals top $1tn

OpenAI has signed about $1tn in deals this year for computing power to run its artificial intelligence models, commitments that dwarf its revenue and raise questions about how it can fund them. Monday's deal with chipmaker AMD follows similar agreements with Nvidia, Oracle and CoreWeave, as OpenAI races to find the computing power it thinks it will need to run services such as ChatGPT. The deals would give OpenAI access to more than 20 gigawatts of computing capacity, roughly equivalent to the power from 20 nuclear reactors, over the next decade. Each 1GW of AI computing capacity costs about $50bn to deploy in today's prices, according to estimates by OpenAI executives, making the total cost about $1tn. The deals have bound some of the world's biggest tech groups to OpenAI's ability to become a profitable business that can meet its increasingly steep financial obligations. "OpenAI is in no position to make any of these commitments," said Gil Luria, analyst at DA Davidson, who added it could lose about $10bn this year. "Part of Silicon Valley's 'fake it until you make it' ethos is to get people to have skin in the game. Now a lot of big companies have a lot of skin in the game on OpenAI," he added. OpenAI is burning through cash on infrastructure, chips and talent, with nowhere near the capital required to fund these grand plans. The deals also involve circular arrangements between the world's most valuable start-up and its partners, as well as complex financing terms that have in most cases yet to be agreed. OpenAI's deals with Nvidia and AMD could cost up to $500bn and $300bn respectively, according to Financial Times calculations, although both include incentives that could also help OpenAI pay for the chips it buys. Oracle's deal will cost OpenAI another $300bn, while data centre group CoreWeave has disclosed computing deals with OpenAI worth more than $22bn. OpenAI also launched an initiative with SoftBank, Oracle and others in January known as Stargate that pledged to invest up to $500bn in US infrastructure for OpenAI. It is not clear how the Nvidia and AMD deal will fit into the Stargate plans. The ChatGPT maker has not disclosed whether it will buy chips directly or through its cloud computing partners, and is expected to lease some Nvidia chips. OpenAI has secured significant financial incentives from its suppliers in return for its chip purchases. Nvidia plans to invest $100bn in OpenAI over the next decade, providing cash OpenAI can use to buy Nvidia's chips for its AI data centres. AMD will give OpenAI warrants entitling it to buy up to 10 per cent of the company for just a penny a share, depending on their project hitting certain targets, including some linked to AMD's share price. AMD shares were worth nearly $204 when markets closed on Monday. If they keep rising, OpenAI could sell its stock to fund its spending on AMD's chips. "It's a pretty innovative structure, which didn't come lightly," AMD chief executive Lisa Su said on Monday. Deals with OpenAI have also given an immediate financial boost to the start-up's partners. Oracle's market value jumped $244bn after its deal was made public last month. AMD shares jumped almost 24 per cent on Monday, boosting its market value $63bn. Their circular relationships have added to concerns about a bubble in AI, at a time when investors are worried that spending on AI data centres is propping up US economic growth. To fund its expansion, OpenAI has raised huge amounts of equity and started to tap debt markets. It secured $4bn in bank debt last year and has raised about $47bn from venture capital deals in the past 12 months -- though a significant chunk of that is contingent on a tricky negotiation with Microsoft, its biggest backer. OpenAI -- valued at $500bn this month -- is also preparing to raise tens of billions of dollars of debt to fund infrastructure, said people close to the company. The start-up's perceived credit risk has prompted concern. Moody's flagged how much of Oracle's future data centre business relies on OpenAI and its unproven path to profitability. OpenAI's arrangement with Nvidia is likely to help investors get more comfortable making large loans. The chip giant, which recently surpassed $4tn in market cap, has routinely used its mammoth balance sheet to invest in companies within its supply chain or among its top customers. The beneficiaries of these deals in turn use the new liquidity to buy more Nvidia chips or borrow money. Nvidia has invested in CoreWeave, which is also a customer and supplier. CoreWeave has also raised raise more than $12bn of debt secured against its Nvidia chips. Beyond Nvidia's support, OpenAI's partners and investors have bought in to its promises of future growth and a path to profitability. OpenAI expects to multiply its revenue from its current level of $12bn in the coming years by rolling out new products and doubling the number of paying subscribers for its core product ChatGPT. Chief executive Sam Altman on Monday said becoming profitable was "not in my top-10 concerns". "But obviously someday we have to be very profitable, and we're confident and patient that we will get there . . . Right now we are in a phase of investment and growth and if we can deliver all of this value," he added. OpenAI and its growing number of partners are betting AI usage will keep growing exponentially. If growth plateaus, or even slows, the investor enthusiasm that has boosted share prices on the back of these deals could quickly falter. One Silicon Valley investment veteran said: "The company is in a far more capital-intensive business than Google or Microsoft ever was, and was born with no cost discipline." Amazon founder Jeff Bezos and Oracle founder Larry Ellison "only found religion" and drastically cut business costs, "after nearly going bankrupt", the investor added.

OpenAI and Nvidia's mega-deals fuel an AI arms race - and fears of a circular bubble

Serving tech enthusiasts for over 25 years. TechSpot means tech analysis and advice you can trust. TL;DR: As rival chipmakers and AI developers weave ever-tighter financial and operational bonds, the pace and complexity of dealmaking shows no signs of slowing. While industry figures celebrate growth and technological promise, the market's interdependent structure is starting to invite doubts. For now, the sector remains gripped by one question: will today's investments transform the economy, or unravel a web of risk whose consequences are yet to be felt? Artificial intelligence investment activity reached new heights this month, as OpenAI cemented strategic partnerships with both Nvidia and AMD, intensifying a trend that has seen billions of dollars cycle between a handful of industry leaders. The recent deals underscore mounting skepticism over the sustainability of the sector's rapid expansion and the intricacies of its business relationships. OpenAI made headlines two weeks ago by signing a deal with Nvidia that could result in up to $100 billion invested in constructing vast data centers designed to fuel the continued growth of advanced AI models. As part of the arrangement, OpenAI committed to purchasing millions of Nvidia chips, raising questions about the nature and scale of their financial ties. The momentum continued on Monday, when OpenAI and AMD announced a multi-billion-dollar partnership in which OpenAI will deploy tens of billions of dollars' worth of AMD chips and become one of the chipmaker's largest shareholders in the process. The channeling of enormous sums into hardware deployment marks a rapid escalation in both spending and ambition for companies whose profitability models remain largely untested. Industry analysts have raised alarms regarding what they refer to as "circular financing" - business arrangements where money often cycles through investments and purchases between closely linked companies, potentially inflating market growth beyond fundamentals. "If we get to a point a year from now where we had an AI bubble and it popped, this deal might be one of the early breadcrumbs," Brian Colello, an analyst at Morningstar, told Bloomberg in reference to Nvidia's investment in OpenAI. Such concerns are amplified by OpenAI's continued cash burn and its expectation that it will not become cash-flow positive until the end of the decade. Nvidia and OpenAI exemplify the pattern, having brokered additional arrangements with companies like Oracle. In a recent announcement, OpenAI confirmed a $300 billion deal with Oracle to build out US data centers using Nvidia and AMD chips. Oracle itself spent billions on Nvidia processors, with financial disclosures showing the company's cloud margins are slimmer than anticipated, heightening doubts about profitability for even established firms participating in the AI infrastructure buildout. The complexity doesn't end with OpenAI and Nvidia. Elon Musk's startup xAI is in the midst of a $20 billion funding round featuring roughly $7.5 billion in equity and as much as $12.5 billion in debt. Nvidia is reportedly preparing to invest $2 billion in the venture, facilitated through a special purpose vehicle that will be used to purchase Nvidia chips. The chips will then be rented out over five years. Similar dynamics are evident with CoreWeave, a newer entrant in cloud infrastructure. Nvidia both invested in and contracted services from CoreWeave, taking a 7 percent ownership stake and promising $6.3 billion in service purchases. OpenAI also received $350 million from CoreWeave before extending cloud agreements up to $22.4 billion. Executives inside the AI sector maintain that these overlapping arrangements are essential to meet extraordinary demand. "It's a virtuous, positive cycle," AMD CEO Lisa Su told Bloomberg TV. OpenAI President Greg Brockman described the environment as requiring "an industry-wide effort" spanning the entire supply chain. Despite industry assurances, some observers see parallels to previous speculative markets. Paulo Carvao, a senior fellow at Harvard Kennedy School who studies AI policy, noted, "In the late 1990s, circular deals were often centered on advertising and cross-selling between startups, where companies bought each other's services to inflate perceived growth. Today's AI firms have tangible products and customers, but their spending is still outpacing monetization." Nvidia, responding to criticism, emphasized that it does not require its investment recipients to use Nvidia products. CEO Jensen Huang told CNBC, "We don't make it a requirement that they use that investment to buy Nvidia's technology. They could use it to do anything they like." OpenAI, on the other hand, declined to comment. Previously, much of the AI sector's funding flowed through established technology giants like Microsoft, Amazon, and Google, enabling strategic alliances that also benefited cloud service providers. Now, newer AI ventures such as OpenAI and xAI are turning to traditional debt markets to finance astronomical infrastructure ambitions. "Altman has the power to crash the global economy for a decade or take us all to the promised land," wrote Bernstein analyst Stacy Rasgon in a recent investor note. CoreWeave CEO Michael Intrator acknowledged circular financing worries, but suggested these concerns will subside as enterprise adoption increases. "When Microsoft comes to us to buy infrastructure to deliver to its clients who are consuming 365 or Copilot, I don't care what the narrative is about circular financing. They have end users that are consuming it."

Why Fears of a Trillion-Dollar AI Bubble Are Growing

For almost as long as the artificial intelligence boom has been in full swing, there have been warnings of a speculative bubble that could rival the dot-com craze of the late 1990s that ended in a spectacular crash and a wave of bankruptcies. Tech firms are spending hundreds of billions of dollars on advanced chips and data centers, not just to keep pace with a surge in the use of chatbots such as ChatGPT, Gemini and Claude, but to make sure they're ready to handle a more fundamental and disruptive shift of economic activity from humans to machines. The final bill may run into the trillions. The financing is coming from venture capital, debt and, lately, some more unconventional arrangements that have raised eyebrows on Wall Street.

Column | AI will trigger financial calamity. It'll also remake world.

"AI could eventually cure cancer, end poverty, and even bring world peace," wrote Dario Amodei, the founder of Anthropic, in a 14,000-word piece entitled "Machines of Loving Grace." "The house of cards is going to start crumbling," Sasha Luccioni, a researcher at the artificial intelligence start-up Hugging Face, told the New York Times recently. "The amount of money being spent is not proportionate to the money that's coming in." The conflicting perspectives are head-spinning in their extremes. On Sept. 22, OpenAI and chipmaker Nvidia announced yet another deal involving sums of money that are vast beyond comprehension. Nvidia agreed to invest up to $100 billion -- about the size of New York City's entire 2024 budget -- in OpenAI and in turn provide it with hard-to-get chips. For believers in Amodei's forecast, the deal added weight to the idea that when it comes to investing in AI, too much is not enough. The announcement added nearly $160 billion to Nvidia's market value, which itself has increased by some $4 trillion over the last three years. But for skeptics, it's another dangerous sign of excess, because such deals, in which players in the AI ecosystem are exchanging money, are simply financial tricks that are camouflaging the evidence of a bubble that is about to burst. "The AI industry is now buying its own revenue in circular fashion," wrote skeptic Doug Kass, who runs hedge fund Seabreeze Partners, in a recent piece. It's precisely what happened in the dot-com boom -- except then we called it "roundtripping." Or at least, we called it that after the crash. But the consequences this time are much, much bigger. As Neil Dutta, the head of economic research at Renaissance Macro, first pointed out in a tweet, the capital expenditures made by a small number of companies are now contributing more to GDP than the spending of all of America's consumers. "Wonder if the black holes ultimately consume each other as well as our economy and power grid in the process?" Kass asked. To put it differently, one of the most notorious calamities of the dot-com era was the collapse of Pets.com. Its peak market value was about $300 million. Since the release of ChatGPT in 2022, the value of America's stock market has risen by $21 trillion. In a recent piece, Michael Cembalest of JPMorgan wrote that since that time, AI-related stocks have accounted for 75 percent of S&P 500 returns, 79 percent of earnings growth and 90 percent of capital spending growth. If the parallels to the dot-com boom do hold true -- and as another skeptic, Harris Kupperman, the founder of hedge fund Praetorian Capital says, "history usually rhymes, especially financial history" -- then the answer might be that both Amodei and Luccioni are right. Two things can be true at the same time. Bets on AI are going to cause financial calamity. And AI is going to change the world. Ah, the first dot-com boom. How innocent it all seems now: Everyone, from start-up CEOs to retail investors, believed they, too, were playing their part in changing the world and, of course, getting rich. A key part of the story was infrastructure. Investors poured tens of billions into new telecom companies such as Global Crossing and Qwest, which were building cross-country fiber networks. So-called CLECs, or competitive local exchange carriers, such as ICG Communications, built modern networks in cities to compete with the stodgy old providers of basic phone services. The stocks of companies such as Lucent and Nortel, suppliers of telecom gear, soared, up tenfold over the course of what was, in retrospect, a giant bubble. Part of what blew the bubble ever larger was layers of financial engineering. Qwest and Global Crossing CLECs started to engage in "dark fiber swaps," which deserve their slightly sinister name. Carriers would sell each other fiber they weren't using yet, with each booking the capacity they had "sold" as revenue, while accounting for the cost of the capacity they'd bought capacity as an asset on their balance sheet, not as a hit to income. In other words, revenue magically appeared, while costs were buried. The former director of European Access Management for Global Crossing, who worked on these deals back in the day, tells me that originally, the agreements had a business purpose: letting a carrier quickly expand into new routes without digging. "That quickly changed when sales/marketing and ultimately the [senior] leadership saw how these deals could massively goose the numbers and close "revenue" gaps," he wrote to me. (His current employer won't allow him to use his name in the press.) As for Lucent and Nortel, they extended multibillion-dollar vendor-financing lines to customers such as Qwest, Global Crossing and ICG, effectively paying their own customers to buy gear. Everyone was able to book revenue and keep the party going as long as the markets remained accommodating -- but no one could foot the bill because in the end, customers didn't materialize, at least not in the time frame that was needed. The unraveling was brutal. Global Crossing, Qwest and ICG all went bankrupt; an SEC complaint against Global Crossing says the dark fiber swaps "yielded no actual revenue" once the round trip was taken into account. When their customers defaulted, the Lucents and Nortels had to take huge write-offs for profits that had been reported but never realized. All their stocks were decimated. The Nasdaq ultimately lost 75 percent of its value from peak to trough. The AI build-out is also financed through a web of interlocking relationships. Increasingly, key details about how the contracts truly work are often elided. For instance, Microsoft, Meta, Amazon, Alphabet and Oracle account for a massive percentage of Nvidia's revenue, as they buy Nvidia's AI chips to build the massive data centers that are needed to power the AI boom. Nvidia, in turn, has spent billions investing in start-ups that buy its chips, including OpenAI and a company called CoreWeave. In fact, at several points, including when CoreWeave's IPO almost failed, Nvidia's customers would have faltered without its support. For their part, Microsoft, Google and Amazon have poured nearly $20 billion into Anthropic and OpenAI alone. Since start-ups' biggest cost is cloud computing, the majority of that cash will probably come back to these investors in the form of cloud revenue. The companies say the deals aren't linked. For instance, Nvidia says its investment in OpenAI would not be used for any "direct purchases" of Nvidia chips. But what does that actually mean? The details of the contracts aren't public, hence the worry. "Global Crossing is Reborn," wrote Kupperman, in a recent piece. There is also an accounting issue that is boosting reported profits. The rate at which a company depreciates an investment affects its income, because the more quickly you have to write something off, the less your profits are. Well-known short seller Jim Chanos has argued that the true economic life for cutting-edge Nvidia chips is two to three years. But companies that buy the chips are often depreciating them over much longer periods. In fact, in early 2025, Meta changed its accounting to lengthen its depreciation, thereby boosting its reported profits. That, too, reminds Chanos of the telecom boom. Nvidia is like Lucent or Nortel, because it's booking all the revenue, and its customers like Meta are like the old telecom companies, which wrote off their expenses slowly. "It's tremendously steroidal for corporate earnings," Chanos says. There are all sorts of counters to this way of thinking -- and there are counters to the counters. One is that there's a legitimate business case for this circular web of financing. By taking equity stakes, companies like Nvidia and the cloud hyperscalers are giving AI start-ups access to far cheaper capital than they could secure on their own. That makes it more likely they will succeed, and reduces the risk that promising projects stall for lack of credit. Then again, just as with the dark fiber swaps in the dot-com days, there is often a legitimate business case. At first. Another is that one of the overlooked triggers of the telecom bust was a technological leap that blindsided the industry. As carriers were racing to lay new fiber, a breakthrough that made each line exponentially more powerful, multiplying existing capacity, began to hit the market. The result was brutal arithmetic: U.S. carriers were on track to boost capacity nearly seventyfold over three years, but suddenly, there was far less demand, because you could do so much more with less. So far, no one credible is predicting an overnight step change that would suddenly render today's AI build-out worthless. Then again, even though the technology that force-multiplied bandwidth was available as early as 1996, no one was able to see the impact the last time, either -- in part because the financial engineering masked the signs of weakness. Unlike the dot-com era's debt-soaked companies, the hyperscalers are mostly funding themselves out of their mountains of cash. Morgan Stanley has estimated that nearly half of the almost $3 trillion ($3 trillion!) it predicts in data-center costs between now and 2028 will be funded by the giants' internal cash generation, not fickle capital markets. Then again, another part of the funding for the data centers is coming from complicated financing arrangements created by the private equity complex. There's too little transparency to easily spot the canary that will signal that the money is about to stop flowing. And even the wealthiest of companies can't spend at this level forever. Kupperman says that even the giants like Alphabet are now spending all of their cash flow on the build-out -- and they're beginning to raise debt. According to his analysis, the AI industry will have to generate $480 billion in revenue just to make a decent return on 2025's spending alone. "There just isn't enough revenue," he says. "There can never be enough revenue." He argues that once shareholders realize the returns won't be there, and the days of huge profitability for companies like Meta and Amazon are done, the stocks will plunge, instead of soaring every time the companies announce that they are spending more. "Eventually shareholders will hate the capital destruction -- even if at first, they cheered it on out of ignorance," he says. After he posted his piece, people throughout the industry, from private equity investors to those inside AI start-ups, reached out to him. "Everyone knows that the numbers don't work, but no one knows that everyone knows this," he says. The true parallel between then and now is also the most worrisome one for investors. If you went back in time to 1998, the internet writ large, from the advent of the iPhone to Netflix to social media to the cloud, was so much bigger and more transformative than even the biggest optimist would have forecast. It was real. It was beyond real. Much of that fiber ended up being used -- although it took decades. And yet, the collapse was devastating, and the casualties were not what most people had foreseen, and the fortunes were made in places, like Google, that no one had predicted. Or as Sam Altman says, "I do suspect that someone is going to lose a phenomenal amount of money." In that sense, even some skeptics are also believers. Says Kupperman about Mark Zuckerberg: "He's thinking about owning the operating system for the world. We're going to have one or two, and if you end up with the winning one, you get a percentage of the profits on everything that happens in human activity. I understand why people are willing to go crazy." In other words, there will be enormous societal change, and deeply uneven and hard-to-predict financial outcomes. And the financial engineering will mask the signs of collapse, until it doesn't. If only AI could solve that conundrum.

Welcome to the mega-blob: AI firms are fusing into one big lump

Why it matters: The AI world is moving into a new era of corporate entanglement with OpenAI's latest megadeal, a "tens of billions of dollars" agreement with AMD that has OpenAI buying mountains of AMD's microprocessors and taking up to a 10% stake in the firm. How it works: AI's leading companies still compete -- sorta. They also work together at a large scale in increasingly esoteric ways. You could call that an ecosystem. You could also call it, as AI critics have, a shell game. * Either way, the AI business is beginning to function like one giant dollar-eating, energy-sucking entity that makes chips, trains models and sketches utopias to justify its runaway costs. * All this is happening well before the takeoff of a "superintelligence" that always seems to lie just over the next ridge. What they're saying: "We are in a phase of the build-out where the entire industry's got to come together and everybody's going to do super well," OpenAI CEO Sam Altman told the Wall Street Journal as the AMD deal was unveiled. * "You'll see this on chips. You'll see this on data centers. You'll see this lower down the supply chain." Indeed, everyone seems to be singing "Come Together" with Altman and OpenAI. * Nvidia announced a massive deal last month in which the chipmaker plans to invest up to $100 billion in OpenAI in stages, with OpenAI using the money to build data centers chock full of Nvidia systems. * Nvidia also recently cut a deal with Intel to invest $5 billion in the troubled U.S. chipmaker. * OpenAI has pulled in additional billions from Oracle and SoftBank to fund its ambitious Stargate data center project in the U.S., with more billions from the UAE to fund a data center in Abu Dhabi. * These partnerships all follow OpenAI's foundational relationship with Microsoft, forged in the company's early days and restructured last month. Meanwhile, OpenAI competitor Anthropic has taken big investments from both Google and Amazon. * Oh, but OpenAI itself also has a deal for services from Google Cloud. * And Microsoft is powering some of its products with AI from Anthropic. Between the lines: The U.S. government itself has become a stakeholder in the AI mega-blob. * The Biden administration and Congress had already gotten into the business of funding domestic chipmaking via the CHIPS Act. * Then the Trump administration decided that in return for CHIPS grants aimed at helping once-dominant U.S. chipmaker Intel recover its manufacturing capacity, the company should give the U.S. 10% ownership. Flashback: Historically, as with the buildout of railroads in the late 19th century, eras of massive growth and speculation have led to corruption and scandal that then provokes regulation and prosecution. * But today, the government has become one more player in the game, and the fast-dealing, anything-goes climate of Trump's second term makes a Progressive Era-style pushback look unlikely for now. Yes, but: The AI boom's massive dollar amounts and hints of investment circularity, where company A supplies company B with the cash to buy A's products, create their own kind of risk. * Just as AI technology has an "interpretability problem," where researchers can't always understand or explain what AI models are doing or why, the financial engineering behind the technology is also getting harder for most investors to map, track and grok. * That troubles veterans of the dot-com bust and the 2008-9 financial crisis, both of which featured opaque, unconventional financing mechanisms that went haywire and caused a ton of collateral damage. The bottom line: The more entangled AI firms get with one another, the more likely any setback to one will turn into a calamity for all.

AI Investment Is Already So Much Larger Than the Subprime Mortgage Bubble That You'll Physically Flinch

No matter how much is said about the hundreds of billions of dollars hovering over the tech industry's head like an anvil in the "AI bubble," it just won't pop. Not only is it refusing to budge, but it's growing by the day. The AI bubble is getting so bad, in fact, that it's making previous market bubbles look like chump change. According to a new assessment by Julien Garran, a research analyst with the firm MacroStrategy Partnership, the AI bubble is now a staggering 17 times the size of the infamous dot-com bubble, a first-of-its-kind run on tech stocks tied to investor hype over the internet. Even worse, Garran estimates that AI now accounts for over four times the wealth trapped in the 2008 subprime mortgage bubble, which resulted in years of protracted crisis across the globe. In the case of the dot-com bubble, according to macroeconomist David Henderson, major economic catastrophe was avoided as the impact of the stock market rush on US GDP growth was minimal. Unfortunately that isn't the case with AI investment, which now accounts for a massive chunk of our economic growth after years of unfettered hype. Prior to the 2008 financial crisis, meanwhile, bullish investors fed into a doomed property market created by banks to turn high-risk mortgages into a font of cash. Like those toxic subprime mortgages, AI has demonstrated very little long-term value -- at least at this point in its life, Garran notes. The trouble with AI, he told MarketWatch, is you "can't create an app with commercial value as it is either generic [as in video games], which won't sell, or it is regurgitated public domain [as in homework], or it is subject to copyright." He adds that it's also a hard product to market effectively, as one AI startup in New York City is making clear as its subway adverts get covered with hostile graffiti. All the while, Garran says, the cost of AI systems is growing exponentially larger, with rapidly diminishing gains in capability. It's a futile exercise to predict what might finally pop the AI bubble, but one thing's clear: we're already at a point of no return. "To find out whether we have hit a wall we have to watch the LLM developers," the analyst said. "If they release a model that cost 10x more, likely using 20x more compute than the previous one, and it's not much better than what's out there, then we've hit a wall." Absent AI, Garran warns the economy is already slowing to a crawl, and it's only a matter of time before the explosive growth in the tech sector begins to reverse, as it did during the dot-com bubble. Given the sheer size of the threat, it seems the best time for the bubble to burst was yesterday. The second best time, perhaps, is right now.

AI isn't in a bubble -- the cash (and the hype) are real, these analysts say | Fortune

For weeks, all the talk on Wall Street is that the growth of the AI sector must, surely, be unsustainable and that this bubble is due to pop. The record-high price of gold, alone, suggests that a lot of investors want a hedge against an implosion in U.S. tech stocks. Yet some analysts are saying that you should believe the hype. They argue: The biggest cheerleader for AI is, of course, Dan Ives at Wedbush who recently published a note titled "Expecting a Robust 3Q Tech Earnings Season to Match the AI Hype; Popcorn Moment." "The cloud stalwarts Microsoft, Alphabet, and Amazon had very robust AI enterprise demand in the quarter based on our field checks. While some investors continue to question the valuations and pace of this tech spending trend, we believe to the contrary the Street is still underestimating how big this AI spending trajectory is," he told clients. He believes these companies will spend $3 trillion on AI over the next three years. Importantly, that spending isn't coming from debt or VC funding, according to Jan Frederik Slijkerman and Timothy Rahill at ING. They recently published a note examining whether all this AI spending might hurt corporate credit quality and discovered that ... everything is totally fine! "Investments by the largest technology companies [Amazon, Alphabet, Meta, Microsoft, and Oracle] are expected to surpass the US$400bn mark in 2026. ... The investments described above are mind-blowing, given their scale. What is even more striking is that these investments have been funded from operating cash flows," they wrote. "From a debtholder perspective, we are less concerned with a potential mismatch between supply and demand, as the large technology platforms mentioned above have funded their expansion plans from their cash flows," they said. Still, surely stocks are overvalued? The majority of gains in the S&P 500 this year have been driven by a handful of tech companies. That concentration risk could hurt investors if there is a pullback. We aren't there yet, according to Jeff Buchbinder, chief equity strategist for LPL Financial in Boston. "The forward price-to-earnings ratio (P/E) of the S&P 500 has yet to reach dotcom era levels, and in fact remains below December 2020 levels because earnings were depressed coming out of the COVID-19 pandemic," he said in a note today. "So large caps stocks are expensive, lifted by AI-driven technology stocks, but not quite to the extremes of 25 years ago." The economics of AI are much more robust than the dotcom era, he says. "Perhaps the key difference between the broader secular AI growth theme and the dotcom era is that large, AI hyperscalers have mostly funded capital expenditures (capex) with strong internal cash flows, not through AI revenue in singularity or by issuing debt or equity. In comparison, dotcom era spending was broadly funded through massive amounts of 'vendor financing,' which ultimately led to the circular flow of capital that fueled the bubble burst." And even if there is a correction, it won't be too bad, argue Samuel Tombs and Oliver Allen at Pantheon Macroeconomics. They estimate that AI capex boosted U.S. GDP growth by 0.3% points. Even if it all disappeared it would not be enough to tip the U.S. into recession, they say. "Weaker growth is more likely than a recession if the AI boom turns to bust," they said in a note to clients. "The likely hit from the AI boom turning to bust would be a significant drag on the economy, but probably a smaller shock than the bursting of the dot-com bubble in 2000, and an ensuing recession would be far from a forgone conclusion." That comes with a caveat: "It would be more alarming, though, if a reversal of AI optimism led to a broader correction in the stock market beyond AI-linked companies, especially if the hit to households' wealth and confidence tipped the fragile balance in the labor market, leading to a jump in the layoff rate," they said.

Is this an AI boom or bubble? Here's what's really happening

Generative artificial intelligence has proved it can generate code, images, and even quarterly revenue. What it hasn't proved yet is if revenue can soar fast enough to keep pace with the trillion-dollar bets that Corporate America is making on AI. Markets are seeing the opening act of a supercycle, even as skeptics point out how quickly the optimism could buckle. So which is it: an AI boom, or an AI bubble? From Wall Street to boardrooms, the argument is playing out on a split screen. One side points to real cash spitting out of the picks-and-shovels layer -- semiconductors, memory, networking, cloud services -- and to customers paying for AI as part of the stack. AI isn't a party trick, proponents say. The other side flags a widening gap between promised capacity and proven monetization: data centers booked years out while businesses can't get past governance reviews, redacted data sets, and workflows that never scale beyond the lab (even as backlogs swell and the power needs sprawl). One way to read this moment is to separate physics from finance. Physics is the substations, transformers, and racks -- the stuff you can trip over in steel-toed boots. Finance is the multiyear contracts, forward guidance, and story momentum -- the stuff that looks like money until it isn't. When physics and finance align, you get an industrial buildout that pays for itself. When they don't, you get a hard slog -- no fireworks, but plenty of friction, defined by the kind of careful phrasing on earnings calls that sends analysts back to the transcript with a highlighter ("sequencing of deployments," "customer readiness," capex phasing," and "revenue recognition in outer quarters"). Even Sam Altman, the OpenAI CEO who has every reason to stay on-message, has warned that investors are "overexcited." He told an audience in August 2025 that markets are racing ahead of reality -- and that the hangover always comes when money floods in faster than the technology can justify. "When bubbles happen, smart people get overexcited about a kernel of truth," he said -- and in his view, AI is in exactly that phase. As signals go, his comments are closer to a siren than a soothing hum. The academic side is equally blunt. A study from MIT researchers found that 95% of corporate generative AI projects hadn't generated profit, a statistic that lands hard against the market's trillion-dollar expectations. Two companies capture the boom-or-bubble split better than any chart. Nvidia proves there's real money in AI now, with record data-center revenue ($46.7 billion in Q2 2025 revenue) and guidance that's turned once-surreal numbers into baseline expectations -- even if growth showed its first signs of cooling after an unprecedented run. Oracle, meanwhile, shows how frothy the future bets can get -- a backlog inflated by a reported five-year, $300 billion deal with OpenAI that doesn't even start until 2027 and accounts for much of its sudden revenue surge. Depending on which story you emphasize, AI looks like a once-in-a-generation boom or a bubble waiting for a pin. Together, they show the tension at the heart of the debate -- one company cashing in now, another priced on tomorrow's promises. The technology works, but the monetization curve may not match the speed of the capital pouring in. Through that lens, the picture sharpens. Chipmakers' blowout quarters make the cash-now case vivid. Clouds that attribute a not-just-noise share of growth to AI that make adoption sticky. Meanwhile, capacity reservations measured in the hundreds of billions keep the careful-now case visible. The answer lives in the seams: capital turning into revenue, costs bending into sustainable pricing, ambition clearing the last mile of adoption. Leadership is narrow. Flows crowd into the obvious beneficiaries. That concentration turns narrative into market plumbing. When a single leader blinks -- slower AI contribution, heavier depreciation, a guidance line that shifts from "accelerating" to merely "healthy" -- benchmarks wheeze. Fragility here isn't proof of a bubble. It's a reminder that a lot of money is riding on a small number of disclosures. That pull-forward changed the definition of normal: more megawatts, more memory, more of everything that makes a modern data center hum. Scarcity made delivery dates the product: If you could take shipment of top-tier accelerators, you could sell them. Scale followed as clouds rolled out specialized AI infrastructure across their fleets; several platforms attributed a visible, double-digit share of cloud growth to AI services -- conversion you can circle in a transcript. Scrutiny is where we are now. Buyers interrogate total cost of ownership, latency, reliability, and governance. Inside enterprises, pilots sort into two buckets. One set actually changes workflows and shows up in the P&L, and the other, larger, set dazzles in demos and fizzles in production. This cycle can overshoot -- capital-heavy cycles often do -- but it isn't built on clicks and wishful CPMs. That's a flywheel, not a fairy tale. There's also path dependence. Once an enterprise builds the data pipelines, security wrappers, and training capacity that's required for AI-infused workflows, that enterprise is reluctant to unwind them. The standard isn't perfection -- it's "good enough at a price that makes sense." As use cases push past novelty -- copilots that reliably save hours, support agents that cut resolution time, internal search that actually finds the right doc -- the spend gets stickier. The proof looks like growth measured in seats expanding without margin erosion, not a slide deck or demo. If software makers can point to AI features that expand usage without cannibalizing existing lines -- and show stable or improving gross margin -- the kind of proof that shows up in filings, not in keynotes. There's also concentration risk hiding in plain sight. With leadership crowded into a short list of names, guidance phrasing from a handful of companies can swing entire stock indexes. When those disclosures lean into "sequencing" or "customer digestion," the market hears "delay" -- and the repricing is instantaneous. Policy piles on. In Europe, obligations for general-purpose model providers are phasing in, lengthening cycles, and raising fixed costs. Inside enterprises, the last mile still bites. Data hygiene, security integration, and workflow redesign are hard -- and until those muscles are common, chief financial officers will keep funding foundational infrastructure -- data, security, pipelines -- while trimming experiments. Bottlenecks slow application revenue, even if the core infrastructure keeps humming. Then comes mix, where storytelling ends and all of that math begins. It's not enough for cloud revenue to be up. AI services need to lift average revenue per customer without bleeding gross margin. Durability is the last bridge. Backlogs and capacity reservations only become money when invoices are cut and cash lands on something that looks just enough like Wall Street's timetable. The more a backlog leans on one or two counterparties, the more binary the cadence becomes. On prices, behavior beats list rates. If last-gen accelerators require permanent markdowns to move, or if "AI included" becomes the only way to sell a suite, margins are being sacrificed to keep adoption steady. Conversely, if vendors can point to AI features that expand seats or drive sustained usage without chewing through gross margin -- and do it quarter after quarter -- then the price curve is doing its job. Policy milestones matter because they change timelines. In Europe, obligations for general-purpose model providers are phasing in, lengthening cycles and raising fixed costs. Export-control shifts reshape who can buy which chips, which feeds straight back into mix and margins. And don't ignore the boring stuff: utility approvals, interconnect go-lives, substation energizations in key regions. If cloud AI contributions hold, if backlogs convert on time, and if unit costs keep sliding, the boom case will feel obvious in hindsight. If recognition slips and power and policy frictions stack, the digestion will show up in equities long before it shows up in substations. Either way, the racks, transformers, and fiber aren't going anywhere -- which is why this debate is ultimately about conversion and cadence, not faith.

The AI boom's reliance on circular deals is raising fears of a bubble

Nvidia plans to invest in OpenAI, which is buying cloud computing from Oracle, which is buying chips from Nvidia, which has a stake in CoreWeave, which is providing artificial intelligence infrastructure to OpenAI. The AI boom that is revolutionizing how people live and work has become increasingly fueled by just a handful of companies turning to one another for the vast amounts of capital and computing power needed to drive their breakneck growth. Some of those partnerships are worth up to hundreds of billions of dollars. Taken together, they have enormously increased the companies' values, helping send U.S. stock indexes to new highs. But as AI investing grows more insular, there is also a risk that the money flowing between these companies is creating a mirage of growth. If the trend accelerates, some analysts warn, a weak link could threaten the viability of the whole industry -- leaving an outsized mark on the U.S. economy. "The experience of a quarter of a century ago [when the dot-com bubble burst] won't necessarily be repeated, but the scale of recent investment increases by tech firms already indicates that they are taking significant risks," analysts with Oxford Economics research group wrote in a recent note. If it starts to become clear that AI productivity gains -- and thus the return on investment -- may be limited or delayed, "a sharp correction in tech stocks, with negative knock-ons for the real economy, would be very likely," they wrote. The latest example of that network of investments came Monday, when OpenAI -- the maker of ChatGPT -- announced a deal with artificial intelligence chipmaker Advanced Micro Devices, or AMD. Under the terms of the OpenAI-AMD partnership, OpenAI will purchase AMD's chips for an undisclosed sum in exchange for the right to take a stake of as much as 10% in the semiconductor giant. The announcement came just weeks after Nvidia unveiled a deal under which it pledged to invest up to $100 billion in OpenAI. "Excited to partner with AMD to use their chips to serve our users!" OpenAI CEO Sam Altman wrote on X. AMD and Nvidia are direct competitors. An Nvidia spokesperson did not immediately respond to an inquiry about whether any OpenAI funds would be used to finance buying its competitor's chips.

Boom or bubble: How long can the AI investment craze last?

San Francisco (United States) (AFP) - The staggering investments in artificial intelligence keep coming: Last week, AI chip giant Nvidia announced it would invest $100 billion to help OpenAI, the frontrunner in generative AI, build data centers. How are these enormous sums possible when the returns on investments, at least for now, pale in comparison? Huge investments AI-related spending is soaring worldwide, expected to reach approximately $1.5 trillion by 2025, according to US research firm Gartner, and over $2 trillion in 2026 -- nearly 2 percent of global GDP. Even though tangible returns fall short of the investments going in, the AI revolution appears unstoppable. "There's no doubt among investors that AI is the major breakthrough technology" -- on par with harnessing electricity, said Denis Barrier, head of investment fund Cathay Innovation. Silicon Valley's mindset "is more about seizing the opportunity" than worrying about any risks, he said. Geopolitical tensions are helping drive the frenzy, primarily to build massive data centers housing tens of thousands of expensive chips that require phenomenal electrical power and large-scale, energy-hungry cooling. From 2013 to 2024, private AI investment reached $470 billion in the United States -- nearly a quarter in the last year alone -- followed by superpower rival China's $119 billion, according to a Stanford University report. Just a handful of giants are on the receiving end, with OpenAI first in line. In March 2025, ChatGPT's parent company raised approximately $40 billion, bringing its estimated valuation to around $300 billion, according to analysts. 'Circular funding' OpenAI is now the world's most valuable company, surpassing SpaceX, worth $500 billion in a deal for employees to sell a limited number of shares. The company led by CEO Sam Altman sits at the center of an AI investment bonanza: It oversees the Stargate project, which has secured $400 billion of the $500 billion planned by 2029 for Texas data centers spanning an area the size of Manhattan. The White House-backed consortium includes Softbank, Oracle, Microsoft and Nvidia. Nvidia, which completed over 50 venture capital deals in 2024 according to PitchBook data, is often chided for practicing "circular funding" -- investing in startups that use the funds to buy its chips. Some analysts criticize this as bubble-fueling behavior. The OpenAI deal "will likely fuel those concerns," said Stacy Rasgon, a Bernstein Research analyst. In the first six months of 2025, OpenAI pulled in around $4.3 billion in revenue, specialist outlet The Information reported this week. Therefore, unlike Meta or Google with substantial cash reserves, OpenAI and competitors like Anthropic or Mistral must be creative in their search for funds to bridge the gap. For AI believers, an explosion in revenue is only a matter of time for a company whose ChatGPT assistant serves 700 million people -- reaching nearly 9 percent of humanity less than three years after launch. 'Up in smoke' Nothing is certain, however. Feeding AI's computing appetite will cost up to $500 billion annually in global data center investments through 2030, requiring $2 trillion in annual revenues to make the expenses viable, according to consulting firm Bain & Company. Even under optimistic assumptions, Bain estimates the AI industry faces an $800 billion deficit. OpenAI itself plans to spend over $100 billion by 2029 -- meaning turning a profit is still a ways off. On the energy front, AI's global computing footprint could reach 200 gigawatts by 2030 -- the annual equivalent of Brazil's electric consumption -- half of that in the United States. Despite the daunting figures, many analysts remain optimistic. "Even with concerns about a possible 'AI bubble'... we estimate the sector is in its 1996" moment during the internet boom, "absolutely not its 1999" before that bubble burst, said Dan Ives, a Wedbush Securities analyst. Long-term, "many dollars will go up in smoke, and there will be many losers, like during the internet bubble, but the internet remained," said the Silicon Valley investor.

The biggest sign yet of an AI bubble is starting to appear

Why it matters: That debt -- and how it is getting structured -- is "almost an acknowledgement that this is getting out of hand," Dario Perkins, managing director of global macro at TS Lombard, tells Axios. What they're saying: Regarding returns on AI expenditures, the Big Tech firms "say they don't care whether the investment has any return, because they're in a race...Surely that in itself is a red flag," Perkins says. * He sees two major issues: increased leverage to fund costly AI infrastructure and few opportunities to make money once that infrastructure is built and paid for with debt. Zoom out: Big Tech is turning to private debt markets and special purpose vehicles. The catch? That kind of borrowing doesn't have to be reflected on balance sheets. * "SPVs mean companies like Meta do not need to show the debt as their debt," Perkins writes in a note. He likens today's financing tactics to the subprime era when firms shifted risk off the books to reassure investors. * Meta seeks $29 billion via private capital for its AI data center buildout. * Other tech giants are tapping the public market for debt. Oracle recently issued $18 billion in debt to fund its AI and infrastructure expansion. Yes, but: Plenty of strategists have reminded Axios of the old Keynesian adage of "the market can stay irrational longer than you can stay solvent." * In other words, this tech-driven bull market could still have legs to create more wealth before the bubble bursts. Perkins, however, isn't convinced. * "I wouldn't touch this stuff now," Perkins says, adding that comparing this market with the dotcom bubble, "we're much closer to 2000 than 1995." Between the lines: Why are tech companies spending this much to win the AI race if the bubble risk is so prescient? * Because the market is rewarding them even if it "makes no economic sense to spend at this level because there's no way they can recoup the value of the capital spending," investor and author Paul Kedrosky notes on the Plain English podcast. * He is also watching how companies are moving financing off the balance sheet: "That for me is a reflection of not wanting the credit rating agencies to look at what they're spending." What we're watching: Hidden debt, recycling of investment, and insider selling are examples of the warning signs of a late cycle, Perkins says. * Perkins doesn't see the economy as tied up in AI like some other macro strategists have argued. That means investors who are well diversified across the U.S. economy and globally could still benefit even if the AI bubble bursts. The bottom line: If hugely profitable tech companies need to mask their borrowing to fund AI spending, it signals they're not confident that they'll soon get the returns needed to justify such investments. That suggests the very spending powering today's earnings boom can't last forever.

Experts Say This Is a Blinking Warning Sign of AI Bubble

The AI gold rush rages on. Multibillion dollar AI deals are being inked left and right between the heavyweights of the tech sector, dazzling us with awesome sums and promising that this revolutionary tech is just getting started. But there's something very worrying about many of these deals: they're often "circular," as a slew of recent coverage has noted, meaning that AI companies are pouring money into one another, creating an illusion of a robust ecosystem that skeptics worry could quickly come crashing down. And many of the deals tie back to Nvidia, the chipmaker whose hardware is underpinning our age of AI, for which it has become the world's most valuable company. Experts warn that all this inter-industry dealing may be one of the most ominous signs of an impending bubble. One recent example is Nvidia's agreement to invest up to $100 billion in OpenAI, as the ChatGPT maker expands its empire of data centers that will demand enough energy to power millions of homes. As part of that deal, announced two weeks ago, OpenAI will in return use Nvidia's chips to fill out its AI facilities. OpenAI announced it had reached a staggering deal with Oracle, the Larry Ellison-led software giant, to buy $300 billion worth of its cloud computing power over the next five years. Oracle already uses Nvidia chips to power some of its cloud computing options, and it reportedly committed to spending another $40 billion on buying more Nvidia chips to supply OpenAI's colossal data center in Abilene, Texas. Perhaps the clearest example yet came this week from fellow chipmaker AMD. On Monday, it announced plans to sell billions of dollars worth of computing capacity to OpenAI. And OpenAI, in exchange, would receive a ten percent stake in AMD for just one cent per share, an extremely advantageous deal. Of course, that's basically just one step removed from AMD outright paying OpenAI to buy its products -- which does not evince a sustainable model for the long term future. The deal, in other words, is circular, as are the ones mentioned previously. If the money keeps circling back on itself, is this really an industry that will be as profitable as the hundreds of billions of dollars of investment suggest? Perhaps not. Circular deals -- or vendor financing -- defined another tech boom that eventually went bust: the dot-com bubble. "In the late 1990s, circular deals were often centered on advertising and cross-selling between startups, where companies bought each other's services to inflate perceived growth," Paulo Carvao, a senior fellow at the Harvard Kennedy School, told Bloomberg. "Today's AI firms have tangible products and customers, but their spending is still outpacing monetization." Nvidia is especially guilty of propagating this ever-widening web of circular agreements, investing billions into AI companies that also happen to be its biggest customers, like OpenAI. It's also reportedly planning to expand its existing investment agreement into Elon Musk's xAI up to $20 billion. xAI already uses tens of thousands of Nvidia chips, and will use some of the cash from the Nvidia investment to buy even more Nvidia chips, Bloomberg reported -- which xAI will then use to rent out to other AI firms. Earlier this year, Nvidia bought a seven percent stake in CoreWeave, a "neocloud" business that rents out access to AI chips. Recently, Nvidia agreed to buy $6.3 billion worth of cloud services from CoreWeave -- cloud services that are, by the way, powered by Nvidia's own chips. One of CoreWeave's customers? OpenAI, which it has agreed to supply $22.4 billion in data center capacity to. With all this cash trading hands, how does anyone make any money? A reality check came this week when The Information reported that Oracle only made a net profit of $125 million in the second quarter quarter off of $900 million in sales from its Nvidia cloud business -- a measly 14 percent margin. That's less profit than many nontech retail companies, the reporting noted. After the news, Oracle's stock slid by 3 percent. These concerns haven't stopped OpenAI from clinching investment deals worth around $1 trillion, despite the fact that it's yet to turn a profit. Many doubt that it will be capable of making its investors' money back. "OpenAI is in no position to make any of these commitments," Gil Luria, an analyst at DA Davidson, told the Financial Times. "Part of Silicon Valley's 'fake it until you make it' ethos is to get people to have skin in the game. Now a lot of big companies have a lot of skin in the game on OpenAI."

How business leaders can survive a 'phenomenal' AI bubble | Fortune

In August, OpenAI CEO Sam Altman delivered a stark message at a dinner in San Francisco: We are in an AI bubble, and it's going to pop. "A lot of people are going to lose a phenomenal amount of money," Altman told journalists. He likened it to the dotcom crash, but potentially worse. A few weeks later, Altman's board chair and fellow AI founder Bret Taylor backed him up: "It is both true that AI will transform the economy [and that] we're also in a bubble." A seemingly endless amount of capital has whipped up the froth for both tech incumbents and startups, which are receiving eye-popping valuations faster than ever. Lots of these companies will implode. But the underlying technology will keep improving regardless of any short-term bubble, and some companies will emerge as the Googles and Amazons of the AI era. One of the fastest-rising startups today is Ramp, operating in the oh-so-unsexy corporate finance and credit card space. It was founded in 2019 by Karim Atiyeh and Eric Glyman, a duo who had sold their first fintech company to Capital One in an eight-figure deal. With Ramp, they set out to create a unicorn -- a startup with a $1 billion valuation -- in two years. No New York-based startup had ever achieved that. Yet sure enough, by the time Ramp hit its second anniversary, it was valued at $1.6 billion. This summer, that number shot up to $22.5 billion. Ramp also announced (exclusively in Fortune) a new milestone: It had surpassed $1 billion in annualized revenue. Those numbers are startling, but so is that multiple. It's a testament to both Ramp's underlying business (it is used by 45,000 companies, though relatively few of the Fortune 500), and to the current AI frenzy, fueled by investors' fear of missing out. For more on how the founders have positioned Ramp as an AI-powered startup that's built to last, read our cover story by Leo Schwartz. And for behind-the-scenes insights from Glyman, subscribe to my podcast, Fortune 500: Titans and Disruptors of Industry, which features Ramp on one of the most recent episodes. (Find me on Apple, Spotify, or YouTube.) While the money is flowing for startups, the need for incumbents to capitalize on AI has become existential. Intel, the once-dominant chipmaker, is in a fight for its life that requires it to aggressively move into leading-edge chips, including those optimized for AI, while drumming up the funding to float that expensive transition. The Trump administration recently bought an unprecedented $8.9 billion stake in Intel, and Nvidia has committed $5 billion to get the business off the ground. Fortune's Geoff Colvin and Lila MacLellan spoke to current and former Intel executives and employees to get their take on what could be the company's last stand.

Top analyst on concerns about Nvidia fueling an AI bubble: 'We've seen this movie before. It was called Enron, Tyco' | Fortune

A top Wall Street analyst has sounded an alarm over the U.S. equity bull market, warning that its remarkable run is built on a precariously narrow foundation: a surge in spending on, and optimistic assumptions about, infrastructure for artificial intelligence (AI). This spending has fueled a boom in the shares of most of the so-called Magnificent 7 and a few dozen related businesses, which have now come to account for roughly 75% of the S&P 500's returns since the rally of the last few years began. The commentary on September 29 by Morgan Stanley Wealth Management's chief investment officer, Lisa Shalett, frames the current market boom as a "one-note narrative" almost entirely dependent on massive capital expenditures in generative AI, raising questions about its durability as economic and competitive risks start to mount. Shalett's critique came squarely in the middle of some people in the AI field -- and many financial commentators around Wall Street -- fretting at market exuberance and beginning to talk openly about a bubble. In an interview with Fortune, Shalett said she was "very concerned" about this theme in markets, saying her office had broadened from a belief that the market would only bid up seven or 10 stocks to roughly 40. "At the end of the day ... this is not going to be pretty" if and when the generative AI capital expenditure story falters, she said. Shalett said she's worried about a "Cisco moment" like when the dotcom bubble burst in 2000, referring to the company that was briefly the most valuable company in the world before an 80% stock plunge. [By "Cisco moment" did she mean a whole bunch of circular financing coming back to bite the company? If so, that would be worth adding/briefly explaining.] When asked how close we are to such a moment, Shalett said probably not in the next nine months, but very possibly in the next 24. When you look at the actual spending and the amount of capital coming into the space, "we're a lot closer to the seventh inning than the first or second inning," she said. Shalett's comments centered on several recent multibillion-dollar deals to scale up data-center infrastructure. As notable substacker and former Atlantic writer Derek Thompson recently noted in a post titled "This is how the AI bubble will pop," so much money is being spent to support AI's energy-consumption needs that it's the equivalent of a new Apollo space mission every 10 months. (Tech companies are spending roughly $400 billion this year alone on data-center infrastructure, while the Apollo program allocated about $300 billion in today's dollars to get to the moon from the 1960s to the '70s.) What's more than a little concerning to Shalett is that one company alone, Nvidia -- the most valuable company in the history of the world, with an over $4.5 trillion market cap -- is at the center of a significant number of these deals. In September alone, Nvidia invested $100 billion in OpenAI in a massive deal, just days after pledging $5 billion to Intel (the Intel agreement was tied to chips, not data-center infrastructure, per se). Fortune's Jeremy Kahn reported in late September on significant concerns about "circular" financing, or Nvidia's cash essentially being recycled throughout the AI industry. Shalett sees this as a major concern and a major sign that the business cycle is headed toward some kind of endgame. "The guy at the epicenter, Nvidia, is basically starting to do what all ultimate bad actors do in the final inning, which is extending financing, they're buying their investors." Shalett expanded on her concerns by saying that companies around Nvidia "are starting to become interwoven." She noted that OpenAI is partially owned by Microsoft, but now Nvidia has also made an investment in the startup, while Oracle and AMD each have their own purchasing agreements with OpenAI. But OpenAI also has a data-center deal with tech giant Oracle, with the "bad news," Shalett notes, that this deal is "totally debt-financed." OpenAI also struck a deal in October with chip-maker AMD that allows OpenAI to buy up to 10% of AMD. "Essentially, Nvidia's main competitor is going to be partially owned by OpenAI, which is partially owned by Nvidia. So, Nvidia can 'own' a piece of its largest competitor. It is totally circular and increases systemic risk." When reached for comment, a spokesperson for Nvidia said, "We do not require any of the companies we invest in to use Nvidia technology." Nvidia CEO Jensen Huang discussed the OpenAI investment in an appearance on the Bg2 podcast with Brad Gerstner and Clark Tang on September 25, calling it an "opportunity to invest" and part of a partnership geared toward helping OpenAI build their own AI infrastructure. When asked about the allegation of circular financing in general and the Cisco precedent in particular, Huang talked about how OpenAI will fund the deal, arguing that it will have to be funded by OpenAI's future revenues, or "offtake," which he pointed out are "growing exponentially," and by its future capital, whether it's raised by a sale of equity or debt. That will depends on investors' confidence in OpenAI, he said, and beyond that, it's "their company, it's not my business. And of course, we have to stay very close to them to make sure that we build in support of their continued growth." Shalett said that she and her team were "starting to watch" for signs of a bubble popping, highlighting the deal announced roughly a week before OpenAI struck its $100 billion data-center deal with Nvidia, when it struck another with Oracle worth $300 billion. Analysts at KeyBanc Capital Markets estimated that Oracle will have to borrow $100 billion of that amount -- $25 billion a year for the next four years. "Every morning the opening screen on my Bloomberg is what's going on with CDS spreads on Oracle debt," Shalett said, referring to credit default swaps, the financial instrument that was obscure before the Great Financial Crisis, but infamous for the role it played in a global market meltdown. CDSs essentially serve as insurance to investors in case of insolvency by a market entity. "If people start getting worried about Oracle's ability to pay," Shalett said, "that's gonna be an early indication to us that people are getting nervous." She added that all the indications to her speak of the end of a cycle and history is littered with cautionary tales from such times. Oracle did not respond to requests for comment. Since the October 2022 bear market bottom and the launch of ChatGPT, according to Shalett's calculations, the S&P 500 has soared 90%, but most of these gains have come from a small group of stocks. The so-called "Magnificent Seven" -- including high-profile names like Nvidia and Microsoft -- plus another 34 AI data-center ecosystem companies, are responsible for, as cited by Shalett and separately by JP Morgan Asset Management's Michael Cembalest, about three-quarters of overall market returns, 80% of earnings growth, and a staggering 90% of capital spending growth in the index. Comparatively, the other 493 names in the S&P 500 are up just 25% -- showing just how concentrated the rally has become. The so-called "hyperscaler" companies alone are now spending close to $400 billion annually on capex supporting AI infrastructure, Morgan Stanley Wealth Management calculated. The economic influence of AI capex is now immense, contributing an estimated 100 basis points -- fully one percentage point -- to second-quarter GDP growth, according to Morgan Stanley's research. This pace outstrips the rate of underlying consumer spending growth by tenfold, underscoring its centrality to both market performance and broader economic data. "People conflate AI adoption, which is in the first inning, with the capex infrastructure buildout, which has been going full-out since 2022," Shalett told Fortune. She cited concerns about the prominence of private equity and debt capital coming into play, as that "tends to produce bubbles, because it may be unspoken-for capacity." In other words, people have money to burn and they're throwing it at things that may not pay off. Shalett waved away macro theories about the labor market or the Federal Reserve. "We think that's missing the forest for the trees because the forest is entirely rooted in this one story" about AI infrastructure. Morgan Stanley's bull-case mid-2026 price target for the S&P 500 is an eye-popping 7,200, but Shalett highlights that even the most optimistic outlook admits that risk premiums, credit spreads, and market volatility do not seem to fully account for the vulnerabilities lurking beneath the AI-fueled advance. Shalett's analysis suggests that AI capex maturity is approaching and some possible slowdowns are already visible. For instance, hyperscalers have already seen free-cash-flow growth turn negative, a sign that investment may have outpaced underlying technology returns. Strategas, an independent research firm, estimates that hyperscaler free cash flow is set to shrink by more than 16% over the next 12 months, putting pressure on lofty valuations and forcing investors to demand more discipline in how these funds are deployed. Shalett was asked about data centers' disproportionate impact on GDP throughout 2025, which media blogger Rusty Foster of Today in Tabs described as: "Our economy might just be three AI data centers in a trench coat." The Morgan Stanley exec said "That's what makes this cycle so fragile," adding that at some point, "we're not gonna be building any data centers for a while." After that, it's just a question of whether you crash: "Do you have a mild 1991-92-style recession or does it really become bad?" Bank of America Research weighed in on the semiconductors sector in a Friday note, writing that vendor financing in the space, especially Nvidia's $100 billion commitment to OpenAI, has been "raising eyebrows." Nevertheless, the team, led by senior analyst Vivek Arya, argued that the deal is structured by performance and competitive need, rather than pure speculative frenzy. In an interview with Fortune, Arya explained why he wasn't worried despite the "optics" being pretty obviously bad. "It's very easy to say, 'Oh, Nvidia is giving [OpenAI] money and they are buying chips with that money" and so on, but he argued the headlines are misleading about how much money is actually being spent and the $100 billion sticker price on the OpenAI deal "scared everyone." Noting that the deal has multiple tranches that will play out over several years to come, he said it's not like Nvidia is "just handing a $100 billion check to OpenAI [and saying] you know, go have fun." "Nvidia didn't fund all of it," Arya said of the wider generative AI capex boom. Citing public filings, Arya argued that Nvidia's entire investment in the AI ecosystem is in fact less than $8 billion or so over the last 12 months, not such a large figure after all. And he's still bullish on Nvidia and OpenAI, he added, because he sees them as the winners of this particular story. "We think they are going to be among the four or five ecosystems that come up. It's not like Nvidia is going and investing in every one of those ecosystems, right? They're only investing in one of those five, which is, of course, the most disruptive," that being OpenAI. When asked about his own fears of a bubble, Arya actually sounded a calmer but strikingly similar tune to Shalett. "I'm extremely comfortable with what will happen in the next 12 months," Arya said, "And I have high sense of optimism about what will happen in the next five years. But can there be periods of digestion in between? Yeah." Explaining that this is the nature of any infrastructure cycle, "it's not always up and to the right." In other words, after the next nine months in Shalett's opinion and the next year in Arya's, the data-center buildout endgame could be in play. "When these data centers are built," Arya said, "they are not built for today's demand. They're built with some anticipation of demand that will develop in the next, you know, 12 to 18 months. So, are they going to be 100% utilized all the time? No." Some of the biggest names in tech and Wall Street offered were hedging hard about the possibility of a bubble on Friday. Goldman Sachs CEO David Solomon and Jeff Bezos, both speaking at a tech conference in Turin, Italy, said they were seeing the same patterns as Shalett. Solomon said the massive amounts of spending weren't fundamentally different from other booms and busts. "There will be a lot of capital that was deployed that didn't deliver returns," he said. That's no different from how investment works. "We just don't know how that will play out." Bezos characterized it as "kind of an industrial bubble," arguing that the infrastructure would pay off for many years to come. OpenAI CEO Sam Altman, who got markets jittery in late August when he mentioned the B-word, was asked again to comment on the subject while touring (what else?) a giant new data center in Texas. "Between the 10 years we've already been operating and the many decades ahead of us, there will be booms and busts," Altman said. "People will overinvest and lose money, and underinvest and lose a lot of revenue." For his part, Cisco CEO John Chambers, one of the faces of the dotcom bubble, told the Associated Press on October 3 that he sees "a lot of tremendous optimism" about AI that is similar to the "irrational exuberance on a really large scale" that marked the internet age. It indicates a bubble to him, but only "a future bubble for certain companies. Is there going to be train wreck? Yes, for those that aren't able to translate the technology into a sustainable competitive advantage, how are you going to generate revenue after all the money you poured into it?" When asked whether the size of this potential bubble represents uncharted waters for the economy, especially considering the one-note nature of the long bull market, Shalett said Wall Streeters are always evaluating risk. But putting on her "American citizen hat," she warned about the media consolidation that sees Oracle's founder Larry Ellison also now playing a major role in TikTok (as part of a buying consortium of Trump-friendly billionaires) and Paramount in Hollywood and CBS News in New York (through his son, David Ellison, the media company's new owner). Shalett said she's worried about "groupthink" filtering into the functioning of markets. "That is not something that most of us have experienced in our lifetimes," she said. "You stop factoring in risk premiums into markets, there is no bear case to anything."

Why Wall Street Analysts Say We're Not in an AI Bubble... Yet

Some warn of early bubble warning signs, including big tech's increasing reliance on debt and an IPO frenzy. A spate of unusual deals within the AI ecosystem has recently fueled concern that the AI boom is actually an AI bubble, but some professional market watchers say this isn't 1999 -- at least not yet. OpenAI has vowed to spend hundreds of billions on Nvidia (NVDA) and Advanced Micro Devices (AMD) chips; in exchange, Nvidia will invest in OpenAI, and OpenAI will invest in AMD. Nvidia has invested in cloud providers Nebius (NBIS) and CoreWeave (CRWV), both of which buy its chips, and agreed with the latter to buy all of its unused computing capacity through 2032. These deals are just a few threads in a growing web of entanglements connecting chipmakers, cloud providers and AI model makers. Skeptics warn these circular deals are evidence of an artificial intelligence bubble forming. They argue that Nvidia, through investments in the companies buying and renting its chips, is subsidizing the AI build-out, and artificially overstating the strength of AI demand in the process. But many on Wall Street disagree, including analysts at Bank of America and Goldman Sachs, who cast doubt on the bubble narrative in notes published Wednesday. "We believe the recent concerns re AI financing are highly overstated," wrote BofA semiconductor equity analyst Vivek Arya, who doesn't expect circular deals to account for more than 5% to 10% of the $5 trillion that is likely to be spent on AI by 2030. Though, it's worth noting that by BofA's calculation, OpenAI will have to spend between $500 and $600 billion on infrastructure as part of its Nvidia deal alone; add in its commitments to AMD and Oracle (ORCL), and the startup has agreed to spend about $1 trillion on cloud computing in the coming years. OpenAI's centrality to the recent spate of deals has concerned some onlookers. The ChatGPT maker recently became the world's most valuable startup, valued at $500 billion, and is in the process of becoming a for-profit company. Still, it anticipates burning through $115 billion on the road to profitability, which isn't expected to happen until the end of this decade at the earliest. However, Arya notes that OpenAI, while a big player in the AI boom, is "only one of multiple (5-10) ecosystems including the four large US hyperscalers, Tesla/xAI, multiple sovereign buildouts (Middle East, Asia), 100+ neoclouds that do not require or have disclosed minimal vendor financing." Skeptics have repeatedly pointed to the stock prices of the Magnificent Seven -- Nvidia, Microsoft (MSFT), Apple (AAPL), Alphabet (GOOG), Amazon (AMZN), Meta (META), and Tesla (TSLA) -- as evidence of a burgeoning bubble. The ten largest companies in the U.S., including all of the Mag 7, cumulatively account for a quarter of global equity market value. "This degree of concentration is, in our view, unsustainable," wrote Goldman Sachs analysts in a note on Wednesday, "but this is not the same as saying that we are experiencing a bubble." Goldman identifies three common components of financial bubbles: "rapidly rising asset prices, extreme valuations and rising significant systemic risks driven by increased leverage." While the Mag 7 stocks have soared over the past three years, Goldman notes that their valuations remain modest compared with the lead up to the Dotcom bubble, in part because their stock prices reflect "powerful and sustained profit growth rather than excessive speculation about the future." Another difference between then and now is that the companies fueling the AI boom with their spending have unusually strong balance sheets, cash flows and hugely profitable businesses not tied to AI. That has allowed them to fund their buildouts with existing revenue streams rather than equity and debt, which reduces the risk that a sudden shock to the AI ecosystem halts AI investment, forces a sharp revaluation of related assets, and ripples through the broader economy. Though, Goldman concedes that, while we're not currently in a bubble, a few developing trends suggest we're headed there. Big tech's debt issuance picked up this year as cash reserves dwindled, suggesting systemic risk may be increasing. More companies have gone public to cash in on the AI frenzy, and investors have eaten the new offerings up. "The starting day premiums for new issues has reached an average of 30% in the US," the highest since the Dotcom bubble, according to Goldman.

Boom or bust: AI insiders concerned about Nvidia, OpenAI's circular dealmaking - The Economic Times

A surge in multi-billion-dollar, interconnected AI deals involving Nvidia and OpenAI has sparked fears of a dot-com-style bubble. Critics warn that circular financing may inflate valuations without creating real value. Supporters argue these massive investments are essential to meet booming AI infrastructure and compute demand.A flurry of billion-dollar circular deals in the AI space has analysts drawing uncomfortable parallels with the dotcom bubble. Most of these deals are centred around two industry leaders -- Nvidia and OpenAI, raising concerns that an increasingly complex and interconnected web of business transactions is artificially propping up the trillion-dollar AI boom. Indeed, detractors are pointing out that these deals do not create any actual value. The proponents, however, say these are necessary to meet the surging demand for the new-age technology. The investments seem to be in line with the current demand for AI compute and superintelligence ambitions. However, when an AI bubble -- where circular financing deals artificially pump up AI players' valuations -- bursts, the blast radius can singe the capital markets and the larger economy. Deal frenzy Nvidia and OpenAI have been signing large, and sometimes overlapping, deals with chipmakers, cloud providers, and startups. The US based chipmaker has invested in dozens of AI startups that use its graphic processing units (GPUs). To a lesser extent, the ChatGPT parent has made similar investments in startups that build services on top of its AI models. Early in September, Nvidia agreed to invest $100 billion in OpenAI for 10 gigawatts (GW) of data centre capacity over 10 years, which would be fitted with millions of GPUs from the former. The day after this deal, OpenAI confirmed that it has signed a $300 billion deal with Oracle to build five data centres in the US, which will also feature Nvidia chips. "If we get to a point a year from now where we had an AI bubble and it popped, this deal might be one of the early breadcrumbs," said Morningstar analyst Brian Colello referring to the Nvidia-OpenAI deal. Earlier this week, OpenAI struck a multi-billion-dollar deal with Nvidia's rival AMD for AI chips, slated to be rolled out next year. AMD has also issued a warrant that offers 10% of the company to OpenAI, making it one of the largest shareholders. These three OpenAI deals alone could exceed $1 trillion. The latest deal in this space is Nvidia's $2 billion equity investment in Elon Musk's xAI, part of a $20 billion round that will be structured via a special purpose vehicle (SPV). The SPV will be used to buy Nvidia chips, which xAI would then lease for five years. Another circular arrangement can be seen in neocloud company CoreWeave, which listed on Nasdaq in March. Nvidia holds 7% stake in it and has a $6.3 billion backstop (added) deal for its cloud services, which rent out access to Nvidia chips. Meanwhile, OpenAI took a $350 million stake in CoreWeave ahead of its IPO, and recently expanded its cloud deals with the company to as much as $22.4 billion. OpenAI and its peers earlier used to rely on Big Tech players like Microsoft, Alphabet, and Amazon for investments. Now they are also tapping the debt market to fund their infra ambitions. Also Read: ETtech Explainer: OpenAI deals with AMD, Nvidia spark bubble concerns Riding the wave Both Nvidia and OpenAI have ridden the AI wave to beef up their valuations. While their deals have propped up the shares of the involved parties, they have been criticised for being circular, where the money goes back and forth between the two parties. "If things go bad, circular relationships might be at play," said Morningstar's Colello. "In the late 1990s, circular deals were often centred on advertising and cross-selling between startups, where companies bought each other's services to inflate perceived growth," Paulo Carvao, a senior AI policy researcher at Harvard Kennedy School who worked in tech in the late 1990s, told Bloomberg. "Today's AI firms have tangible products and customers, but their spending is still outpacing monetisation," he added Case in point: OpenAI. The most valuable unlisted company in the world is yet to turn a profit while it plans to invest trillions in AI infrastructure in the race to artificial general intelligence (AGI) supremacy. The $500-billion AI startup is relying on a mix of venture capital, debt, and creative partnerships with businesses to fuel its ambitions. "Obviously, one day we have to be very profitable and we're confident and patient that we will get there... But right now, we're in a phase of investment and growth," OpenAI cofounder and CEO Sam Altman said at a developers' event earlier this week. Nvidia, the most valuable company in the world and the dominant player in AI chips, is also investing in AI at a rapid pace: the company signed 50 deals by September, compared to 52 in 2024, Bloomberg reported citing data from financial research platform Pitchbook. The investments in AI infrastructure are not a bubble, but represent the new normal to meet skyrocketing user demand, Fidji Simo, chief executive of OpenAI's applications, said on Monday. CoreWeave chief executive Michael Intrator acknowledged the circular financing worries in a recent interview with Bloomberg News, but expressed confidence that the concerns will dissipate as more businesses adopt AI. "When Microsoft comes to us to buy infrastructure to deliver to its clients who are consuming 365 or Copilot, I don't care what the narrative is about circular financing," Intrator said. "They have end users that are consuming it." On the flipside, analysts worry that a drop in demand or a Chinese rival coming up with cheaper alternatives could cause the AI bubble to pop. Altman "has the power to crash the global economy for a decade or take us all to the promised land," Bernstein researcher Stacy Rasgon wrote in a recent investor note. "Right now we don't know which is in the cards."

AI Blob: Capitalism's New Feeding Frenzy

Enter your email to get Benzinga's ultimate morning update: The PreMarket Activity Newsletter If you want to see capitalism at its most feverish, forget crypto or meme stocks. The new mania has a more respectable name: AI Infrastructure. Data centers, mazes of silicon and cooling coils, are devouring every spare dollar, pound of copper and parcel of land they can find. In a recent market note, Michael Cembalest, Chairman of Market and Investment Strategy at J.P. Morgan, drew parallels with the cult horror movie "The Blob." The explosion of capital spending by AI players, he wrote, has created a self-reinforcing loop of stock valuations, investor expectations, and capex that's hard to stop once it gets rolling. In charts, he notes, a handful of companies -- the Nvidia, Microsoft, Amazon, and Google -- now account for an overwhelming share of market returns. "AI stock participation has been exceptionally narrow," Cembalest wrote. The S&P 500's breadth is shrinking even faster than it did during the dot-com bubble, he warns. See Also: AI's $1.2 Trillion Growth Will Rest On Data Centers And Power, Not Financing Schemes The Capex Spiral Nobody Can Escape While the original Blob consumed organic matter, the current one consumes electricity. Researchers estimate that the U.S. data center footprint will increase significantly by 2030. It may require as much as 12% of the nation's total electricity generation -- up from around 4% today. For a country already struggling to modernize its grid, that's not just a logistical problem but a geopolitical one. Add subsidies, tariffs, and defense-related funding into the mix. The global race for chips, energy, and computing capacity is starting to resemble a new form of mercantilism. Everyone wants their AI sovereignty, and the U.S. is ensuring its capital and energy remain onshore. That brings us to the other half of the madness: the spending. According to McKinsey, the bill for chips, data centers, and energy required for AI in the next five years is going to be $6.7 trillion. The firm outlined five key investor archetypes participating in the rally, including: Builders Energizers Rechnology developers and designers Operators, and AI architects. Furthermore, McKinsey highlights the shaky success of AI pilot projects in companies. Senior Partner Pankaj Sachdeva puts it at "less than 15%." Then, throw in geography and energy. New hyperscale sites are being established in areas with affordable land and promises of renewable energy -- such as Texas, North Dakota, and New Mexico -- because traditional clusters often lack spare power. However, remote sites incur additional costs (such as transmission build-out and grid upgrades) and risks (including stranded assets and overbuild). While such an enormous bill requires a balance between prudent and rapid capital deployment, perfectly aligning the interests of five parties might be a challenge. "You do not know which of these players will be around in five, ten or 15 years' time," Sachdeva says when discussing risks of utilities signing long-term contracts. AI-powered Ouroboros This is where things look messy. To fulfill its appetite for metals and megawatts, Silicon Valley has found a more circular way to do it: fund itself. The most bizarre part of this story is what Cembalest calls the "circular capital ecosystem," where companies invest in each other's dreams to sustain their own valuations. OpenAI inks massive capacity deals with Oracle, AMD, and Nvidia; Nvidia buys stakes in cloud players; cloud players commit to Nvidia kit; and startups get financed by the very vendors they will buy from. This isn't new -- railroads and telecom booms ran on similar self-referential optimism -- but the velocity today is unmatched. The AI ecosystem has become a loop of spending and market capitalization, where each new round of capital investment is validated by rising stock prices rather than realized profits. The risk, of course, is what happens when returns flatten but depreciation and power bills don't. In the end, the AI boom looks less like a straight line to utopia and more like a high-voltage treadmill. Every layer -- from miners in Chile to cloud giants in Silicon Valley -- is running faster just to stay in place. The capital is circular, the energy is real, and the risk is that the "Blob" consumes more than just voltage. It might consume its creators as well. Read Next: AI Companies Are Burning Billions, Tech Workers Are Getting Wrecked -- And Wall Street Doesn't Care Image: Shutterstock Market News and Data brought to you by Benzinga APIs

Dizzying deal delirium: How the AI bubble bursts | Fortune

So let's get this straight: OpenAI is now taking a 10% stake in AMD, while Nvidia is investing $100 billion in OpenAI; and OpenAI also counts Microsoft as one of its major shareholders, but Microsoft is also a major customer of AI cloud computing company CoreWeave, which is another company in which Nvidia holds a significant equity stake; and by the way, Microsoft accounted for almost 20% of Nvidia's revenue on an annualized basis, as of Nvidia's 2025 fiscal fourth quarter. In less than three years, OpenAI has gone from a parlor game to a pillar of the global economy. You cannot help but ask, "Is this like the Wild West, where anything goes to get the deal done?" One company grants a chip supplier equity for financing data-center buildouts but takes an ownership stake in another manufacturer for the eventual development of a similar product. It is hard to imagine that Jensen Huang outsmarted his equally talented cousin, the AMD CEO, Lisa Su. The lines between revenue and equity are blurring among a small group of highly influential technology companies, to the tune of hundreds of billions of dollars. Su has defended AMD's deal with OpenAI, asserting that market bears are "thinking too small." The chief executive describes AI's potential as sparking a new decade-long "Supercycle" that will "transform industries, from finance to healthcare and research." We have seen this story before, back during the "cable cowboy" days. Programmers paid the distributors. Or wait -- distributors paid the programmers. And of course, there are the frequent comparisons to the run-up to the dot-com bubble. When a dramatic technological change occurs, people are often unsure exactly what to do, but they frequently act as if they do confidently know the best path forward. Major players in the industry have begun to call out the AI euphoria. Just last Friday alone, three of them spoke out, hedging hard on what's to come. Goldman Sachs CEO David Solomon said he expects there to be "a lot of capital that was deployed that [doesn't] deliver returns." Amazon founder and executive chairman Jeff Bezos called the current environment "kind of an industrial bubble." Sam Altman, CEO of OpenAI, warned that "people will overinvest and lose money" during this phase of the AI boom. At our Yale Chief Executive Leadership Institute CEO Summit in June, we heard similar admonitions from over 150 top CEOs, including seasoned venture capitalists, renowned technology founders, and global partners from leading consulting firms. While the commercial outlook for AI among business leaders was enthusiastic, there were significant pockets of concern, extending beyond safety fears, to question the frenzied paths of investment. Sure, 60% of CEOs polled didn't believe that AI hype had led to overinvestment; however, the other 40% raised significant concerns about the direction of AI exuberance, believing a correction to be imminent. Reports estimate that AI-related capital expenditures surpassed the U.S. consumer as the primary driver of economic growth in the first half of 2025, accounting for 1.1% of GDP growth. JP Morgan Asset Management's Michael Cembalest notes that "AI-related stocks have accounted for 75% of S&P 500 returns, 80% of earnings growth and 90% of capital spending growth since ChatGPT launched in November 2022." More concerning, RBC's Kelly Bogdanova points out that after the massive earnings growth of 2023 and 2024, growth rates between the "Magnificent Seven" and the rest of the S&P 500 are expected to converge next year. Notably, she recognizes that "the gap between the Tech sector's share of market cap and net income has widened significantly" since late 2022. At our June CEO Summit, David Siegel, a computer scientist and an early student of AI at MIT, and later a Co-Founder of quantitative hedge fund Two Sigma, candidly advised, "[AI technologies are] transforming business ... but I also believe that the current wave of AI hype continues to mix fact with speculation freely." Siegel continued, "Rarely does anyone speak about the limitations of current AI technologies." The renowned investor and technologist has long held these beliefs but was emboldened by the groundbreaking report from Apple suggesting that the reasoning capabilities of AI models may not be as sophisticated as many assume. Siegel explained in simple language what the findings may mean: "AI researchers have long worried that the impressive benchmarking results [of AI models] may be due to data contamination, where the AI training data contains the answers to the problems used in benchmarking. It's like giving a student the answers to a test before they take the exam. That would lead to exaggerations in the models' abilities to learn and generalize." A recent study from MIT -- released after the June CEO Summit -- backed Siegel's claims. Their research revealed that 95% of the 52 organizations considered had achieved zero return on investment, despite spending $30 billion to $40 billion on GenAI across more than 300 initiatives. With uncommon public candor from the consulting world, AlixPartners Co-CEO Rob Hornby, recognized among business leaders for his expertise in the technology industry, shared a similar view, telling the CEO Summit group, "I don't think [AI models are] ready for sustaining long chains of activity in ways that displace people ... AGI is just not close ... I don't think artificial intelligence and humans have that much in common, right now." Hornby's comments starkly contrast with the supposed reckoning of mass layoffs that the founders of leading AI companies believe will soon come to the labor force. Anthropic CEO Dario Amodei made headlines recently after telling Axios, "AI could wipe out half of all entry-level white-collar jobs -- and spike unemployment to 10%-20% in the next one to five years." Another candid consulting leader, Asutosh Padhi, Senior Partner and Global Leader of Firm Strategy at McKinsey & Company, took a more balanced view of AI in the workforce. He framed the technology as a source of enhanced productivity, not necessarily a mass replacement for people. McKinsey will continue to "hire extraordinary people, [with AI] helping them be even better at what they do," said Padhi. Greycroft founder Alan Patricof, the venture capital pioneer with over 60 years of experience, extended a more nuanced view of AI to the investment community: "The AI revolution is a true revolution ... [but] I am cautious about valuations and what people think can be accomplished in the short term ... A lot of people have run into this field, and just because 'AI' is attached to the name, or they incorporate it into their business plan ... [it] gets a lot of people excited." Pitchbook reported that nearly two-thirds of deal value in the U.S. went to AI and Machine Learning startups in the first half of 2025, up from 23% in 2023. The meteoric rise can be primarily explained by the increased focus of venture organizations, such as Andreessen Horowitz and Y Combinator, on AI startups and the mindboggling valuations of those emerging companies. Under such exuberant conditions, Patricof reflected, "There will be winners and losers, and the losses will be pretty significant." The warnings of exuberance may be mounting, but how the bubble pops is a question that has gone unanswered. The possibilities are endless, but three stand out as having a higher likelihood of occurring. A small group of companies is securing most of the major deals. News about multibillion-dollar investments from familiar companies such as OpenAI, Nvidia, CoreWeave, Microsoft, Google, and a few others is reported almost daily. Should the bold promises of AI fall short, the dependence among these major AI players could trigger a devastating chain reaction, causing a widespread collapse similar to the 2008 Great Financial Crisis. Worse yet, the ambitions are numerous and compounding, with large energy and grid infrastructure buildouts, highly advanced agentic capabilities, and widespread commercial and consumer adoption all anticipated over the next five years. Take one example. OpenAI is committed to investing $300 billion in computing power with Oracle over the next five years, which averages $60 billion per year. Besides losing billions of dollars annually, OpenAI's projected revenues are expected to reach $13 billion in 2025, requiring even larger amounts to cover future shortfalls. The announcement of the deal caused Oracle shares to soar by over 40%, adding nearly one-third of a trillion dollars to the company's market value in a single day. OpenAI's valuation has almost doubled from $300 billion to $500 billion in less than a year. Notably, recent reporting by CNBC suggests that the deal for Oracle may be costly, with the company expecting to "lose considerable sums of money" on its rental of data centers, primarily to OpenAI, and already incurring a $100 million loss in the most recent quarter. Another report says the loss may be a simple timing issue. Not long ago, Sam Bankman-Fried promised to revolutionize financial market operations with cryptocurrency exchange FTX and trading firm Alameda Research. However, poor governance and limited regulatory oversight proved disastrous for Bankman-Fried and his backers when his fraudulent activities were exposed. Nefarious actors using Binance for money laundering shortly after the collapse of Alameda Research set the industry further back. Blockchain technologies do offer promising advances to antiquated sectors and practices, but they must be in the right hands and have the proper guardrails. AI is in a similar position to the cryptocurrency exchanges of the early 2020s, with much to offer but disparate governance practices and minimal regulatory oversight. But back then, the cryptocurrency market was still relatively small and viewed as risky by the average investor, which limited the fallout. The perceived value of AI is exponentially larger, and the potential damage from bad or even questionable actors is, therefore, much greater. Anthropic CEO Dario Amodei, Google CEO Sundar Pichai, and xAI CEO Elon Musk have each raised concerns about the "probability of doom" from AI misuse. Amodei estimates there is a 25% chance that AI will go "really, really badly." Ironically, Musk's Grok, xAI's large language model, provided a recent example of what happens when tampering with the inner workings of AI models goes awry. It is not difficult to imagine a major, publicly available AI model going rogue and inflicting significant damage to financial markets or national security systems. Such an action would require a national moratorium on comparable AI models until the damage is contained and the risk mitigated. In a powerful Washington Post op-ed, Bethany McLean poignantly recalls the overbuilding of fiber-optic cable infrastructure during the 1990s dot-com bubble. Part of the problem was due to circular financial engineering, but the other factor was due to "a [technological] breakthrough that made each line exponentially more powerful, multiplying existing capacity," rendering much of the infrastructure unnecessary for decades. For AI, further innovation in semiconductor chip design or major advances in quantum computing, as hundreds of billions of dollars in data center infrastructure are being deployed, would immediately leave much of that investment useless in the medium to long term. That is not to say that the spare "compute" will not be needed in the future, but as McLean notes, just like the fiber-optic cable infrastructure, it could be years before those data center investments start generating a return for their backers. In the business classic Extraordinary Popular Delusions and the Madness of Crowds, Charles Mackay examined the psychology of crowd behavior and mass hysteria throughout history, from the Dutch Tulip Mania of the 1630s to humanity's historical obsession with transmuting base metals into gold. While Mackay wrote his book in 1841, the AI mania continues to validate his conclusion: "Men, it has been well said, think in herds; it will be seen that they go mad in herds, while they only recover their senses slowly, one by one."

ETtech Explainer: OpenAI deals with AMD, Nvidia spark bubble concerns - The Economic Times

OpenAI is binge shopping for computing clout. In less than a month, the AI champ has signed multiple agreements for chips and cloud capacity of several gigawatts (GW). Considering the rising demand for AI -- in both enterprise and retail -- and OpenAI's hardware ambitions, these deals hold a critical position in the company's infrastructure plans. As per a Financial Times report, the combined value of these deals now stands at $1 trillion, compared to OpenAI's valuation of $500 billion. The deals have, however, raised concerns of circular financing, and analysts are worried these could lead to an AI bubble. Here's a look at OpenAI's recent deals and what they mean. OpenAI deals: Oracle, Nvidia, AMD On September 10, OpenAI and Oracle signed a $300 billion deal for cloud and computing capabilities. The five-year deal is part of the 10GW Stargate AI infrastructure project, which has the companies as partners, along with SoftBank. The three partners announced five new Stargate sites in the US on September 23. Extending the original scope of the project, data centres will also be built in the UK, the UAE, Norway, India, and South Korea. On September 24, OpenAI signed a $100 billion deal with leading chipmaker Nvidia. Nvidia will invest the amount in OpenAI over a period of 10 years -- starting with the first instalment of $10 billion -- which the AI company will use to buy Nvidia's graphics processing units (GPUs). The plan is to deploy 10 GW of Nvidia-powered AI data centres. On Monday, OpenAI and AMD signed a multi-billion dollar deal that sent the latter's stock shooting. Under the deal, AMD will provide OpenAI with 6 GW of the latest version of its high-performance GPUs, starting next year. AMD also issued a warrant that allows OpenAI to buy up its 10% of its shares. The warrant will vest based on two milestones tied to the amount of computing power deployed. Bubble trouble? In case of Nvidia, OpenAI would use the deal proceeds to buy computing power from Nvdia. This amounts to circular financing. As for Oracle and AMD, there are concerns whether the company will be able to honour its commitments given its financials, and what'll be the impact on the industry if that does come to pass. The risks also include demand faltering or a Chinese rival coming up with a cheaper alternative. But for the moment all is sunshine and happiness as the share prices of Oracle and AMD have spiked. Also Read: OpenAI's Sora 2 tops Apple's App Store amid copyright backlash No bubble: OpenAI apps CEO The dizzying investments in AI infra are not a bubble, but represent today's "new normal" to meet skyrocketing user demand, Fidji Simo, chief executive of OpenAI's applications, said on Monday. "What I am seeing here is a massive investment in compute, with us meeting that need for computing power so incredibly badly for a lot of use cases that people want. [Video AI generator] Sora is a great example right now -- there's much more demand than we can serve," Simo told news agency AFP. The recently released Sora 2 short AI video platform has gained traction, with an eye on the market share of TikTok and Meta's Reels.

OpenAI Computing Deals Pass Trillion Dollar Mark | PYMNTS.com

That's according to a report Tuesday (Oct. 7) by the Financial Times (FT), which notes that these agreements dwarf the artificial intelligence (AI) startup's revenues while raising questions about how it can honor those commitments. The report came one day after OpenAI forged a new agreement with chipmaker AMD, following similar deals with Nvidia, Oracle and CoreWeave, as the company seeks the computing power it projects it will need to power AI offerings such as ChatGPT. According to the FT, these deals would give OpenAI access to upwards of 20 gigawatts of computing capacity, around the equivalent of power from 20 nuclear reactors, over the next decade. Each 1GW of AI computing capacity costs about $50 billion to use at current prices, per estimates by OpenAI, making the total cost about $1 trillion. These deals, the report added, have tied some of the world's largest tech companies to OpenAI's ability to turn a profit and meet its heavy financial obligations. The company "is in no position to make any of these commitments," said Gil Luria, analyst at DA Davidson, who added that OpenAI could lose roughly $10 billion this year. "Part of Silicon Valley's 'fake it until you make it' ethos is to get people to have skin in the game. Now a lot of big companies have a lot of skin in the game on OpenAI," he said. In other AI news, PYMNTS wrote earlier this week about the changing face of AI funding. Billions are still flowing into the industry, "but not where they used to," that report said. The largest recent funding rounds went to firms focused on deployment, compute and pricing, the systems that determine AI's efficiency and potential for profitability. "The focus is shifting from invention to execution as investors look for what they can scale," PYMNTS added, citing recent rounds from the likes of Cerebras Systems, which raised $1.1 billion at an $8.1 billion valuation to boost chip production and data center capacity. "The movement of capital toward companies like these reflects a broader transition. Investors are supporting firms that make AI usable, reliable and measurable, turning research advances into systems that work at scale," that report said. "As PYMNTS reported, more than half of global venture investment this year went to AI startups, a trend reflected in this week's funding activity."

OpenAI, Nvidia And Oracle Building A High-Tech House Of Cards? Expert Warns 'Infinite Money Glitch' Could Trigger AI-Led Job Losses And Market Chaos - First Trust DJ Internet Index Fund (ARCA:FDN), Advanced Micro Devices (NASDAQ:AMD)

A critical examination of the AI industry's financial mechanics, dubbed the "AI Infinite Money Glitch," is raising alarms among financial analysts and tech critics. Gordon Johnson, CEO and Founder of GLJ Research, LLC, explicitly warns that the current AI boom is not only a potential "high-tech house of cards" but also an "excuse to eliminate millions of American jobs." AI's Circular, High-Stakes Funding Loop The "Infinite Money Glitch," highlighted by Johnson in a recent analysis by Reef Insights, describes a circular flow of tens of billions of dollars between major players like OpenAI, Nvidia Corp. (NASDAQ:NVDA), and Oracle Corp. (NYSE:ORCL). At its core, cloud computing giant Oracle purchases vast quantities of high-end AI chips (GPUs) from Nvidia and Advanced Micro Devices Inc. (NASDAQ:AMD) to build massive data centers. OpenAI then commits to long-term contracts with Oracle to rent this computing power for its AI models. Simultaneously, OpenAI also strikes direct deals with chip makers like AMD, even securing stock warrants in exchange for future supply. Symbiotic Deals Linking OpenAI, Oracle, And Nvidia "The money flows in a circle," Reef Insights explains, detailing how Oracle committed to buying $40 billion in NVIDIA chips for an OpenAI data center. This is reciprocated by Nvidia's plan to invest up to $100 billion in OpenAI, tied to developing significant NVIDIA systems. AMD similarly benefits, supplying OpenAI with chips totaling up to 6 gigawatts, in exchange for a warrant to purchase 160 million AMD shares -- roughly 10% of the company. Reef Insights argues this intricate web of mutual spending creates a system where "each one's revenue depends on another's spending," with the entire edifice resting on the precarious assumption that "real users will pay enough for AI tools to support the whole system." See Also: Jeff Bezos, David Solomon And Now Sam Altman Warns Of A Brewing AI Bubble, But Expert Says 'They Want The Bubble To Pop:' Here's Why Critics Warn AI Boom Is A Jobs-Killer Gordon Johnson, expressing a starker view on X, stated, "You left out the part where 'AI' is used as an excuse to eliminate millions of American jobs, while also exploding the cost those same Americans pay for electricity... Immigrants taking Americans jobs = bad. AI taking American jobs = good (somehow)." Johnson's comments underscore a growing concern about the economic and social implications of an AI-driven future where job displacement is a significant factor, potentially without a corresponding benefit to the wider populace. How Can This Circle Break? The "circle could break," according to Reef Insights, if "OpenAI doesn't make enough money from its users." A lack of real demand would leave Oracle with underutilized, expensive data centers and halt new chip orders for Nvidia and AMD. This scenario draws parallels to the fiber-optic network overbuilding of the early 2000s, where significant investments led to a bust when anticipated demand failed to materialize. "Companies could be stuck with expensive obligations even as usage drops," the analysis concludes, suggesting a rapid decline in profits and stock prices, transforming the AI boom into an "overbuilt cycle." Price Action Here is a list of some AI-linked exchange-traded funds that investors can consider. The SPDR S&P 500 ETF Trust (NYSE:SPY) and Invesco QQQ Trust ETF (NASDAQ:QQQ), which track the S&P 500 index and Nasdaq 100 index, respectively, rose in premarket on Wednesday. The SPY was up 0.13% at $669.97, while the QQQ advanced 0.16% to $605.45, according to Benzinga Pro data. Read Next: Billionaires Ken Griffin, David Tepper, Philippe Laffont And More Load Up On Nvidia Stock As This Tech Bull Sees NVDA 'Going To $5 Trillion' Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors. Image via Shutterstock AMDAdvanced Micro Devices Inc$213.631.00%OverviewFDNFirst Trust DJ Internet Index FundNot Available-%FTECFidelity MSCI Information Technology Index ETFNot Available-%IGMiShares Expanded Tech Sector ETFNot Available-%IXNiShares Global Tech ETFNot Available-%IYWiShares U.S. Technology ETF$197.58-0.11%MAGSRoundhill Magnificent Seven ETF$64.750.17%NVDANVIDIA Corp$186.170.61%ORCLOracle Corp$285.600.48%QQQInvesco QQQ Trust, Series 1$605.380.14%QTUMDefiance Quantum ETF$110.80-0.02%SPYSPDR S&P 500$670.030.14%Market News and Data brought to you by Benzinga APIs

Dot-com fears rise with tech stocks seeing $100 billion swings | Fortune

Investors are excited about OpenAI's expansion driving big gains in technology stocks, but a rising number of Wall Street pros fear that the wild pops that add tens of billions of dollars in value in mere minutes are signaling an unhealthy market reminiscent of the dot-com era. Advanced Micro Devices Inc. took this rocket ride on Monday, as the company's stock soared, briefly boosting its market capitalization by roughly $100 billion at an intraday high, after the chipmaker signed a deal with OpenAI that could lead to billions of dollars in revenue. AMD's shares extended gains into a second day, rising as much as 7.5% in early trading on Tuesday. This follows a 36% jump in Oracle Corp. shares last month, which added $255 billion to the software firm's market value in a single session, after it gave blockbuster guidance for its cloud business, including an agreement with the ChatGPT operator worth $300 billion over five years. "If any one of these deals falls through it has this domino effect downstream that I think is concerning," said Brian Mulberry, client portfolio manager at Zacks Investment Management Inc., which has about $12 billion in assets. "It reminds me of what happened with telecom back in the mid-nineties." The moves come amid growing concern about a bubble forming around artificial intelligence as the key players -- namely Nvidia Corp. and OpenAI -- pledge billions of dollars in deals with a cohort of companies making infrastructure for the technology. As more money is spent, there's mounting fear that the trend will end in a crash the way it did 25 years ago following the dot-com euphoria, when heavy investments were made in anticipation of internet traffic that took much longer to materialize. An unwinding could be even more painful today, as the top tech stocks account for roughly 35% of the S&P 500 Index, compared with less than 15% in 1999. "The market is pricing these deals as if everyone who transacts with OpenAI will be a winner," said Michael O'Rourke, chief market strategist at Jonestrading. "OpenAI is a negative cash flow company and has nothing to lose by signing these deals. Investors should be more discerning. But this is a buy-first, ask-questions-later environment." Hedge fund billionaire Paul Tudor Jones said the current backdrop reminds him of the dot-com bubble in an interview with CNBC's Squawk Box on Monday. "All the ingredients are in place for some kind of a blow off," he said. "Will it happen again? History rhymes a lot, so I would think some version of it is going to happen again," adding that this environment is "more potentially explosive than 1999." One of the big concerns about the deals is their circular capital structures, with the companies using each other's money to buy each other's products, Mulberry said. In addition, the scale of the stock market moves for such large companies is alarming, he said. These are companies with "very large mature balance sheets that are participating in these types of rallies," Mulberry said. "That is unusual, and it does cause a little bit of reflection." To be sure, AMD's climb may be justified because the OpenAI deal signifies a major step in its progress with graphics processing units, or GPUs, where it competes for market share with Nvidia. Wall Street analysts covering AMD generally cheered the news. "This agreement fundamentally shifts how the industry will view AMD's competitive position going forward," Benchmark analyst Cody Acree wrote in a note to clients on Monday, boosting the firm's price target to $270 from $210. "Beyond the obvious revenue and earning accretion of the agreement, we believe this announcement is a ringing endorsement of AMD's increasingly competitive position as a viable technical alternative to Nvidia's AI GPU dominance." Still, having multiple large technology stocks surge by double-digits in quick succession could be a sign that valuations have become disconnected from underlying fundamentals, and that investors are buying primarily based on the fear of missing out on further gains. "The price discovery is actually pretty scary," Ted Mortonson, a technology strategist at Robert W. Baird & Co., said after Oracle's pop. A company that large, gaining so much market value that rapidly is "not good and normal," he added. "I would call it a part of the exuberance bundle."

AI gold rush: Why experts fear a massive trillion-dollar crash could be coming

The AI boom is driving huge spending by companies like OpenAI and Meta, raising fears of a trillion-dollar AI bubble. Experts warn that profits may not match the hype, and smaller startups face big risks. With massive data centers, high-tech chips, and rising competition from China, the AI market could see major gains or a serious crash in the future. The world is spending record-breaking amounts on Artificial Intelligence. Experts have been warning that the current AI craze might be a speculative bubble -- just like the dot-com bubble of the 1990s which ended in a big crash and bankruptcies. Tech firms are pouring hundreds of billions into AI chips, massive data centers, and tools to keep up with chatbots like ChatGPT, Gemini, and Claude. The total AI spending could reach trillions of dollars. The money is coming from venture capital, loans, and even unusual deals that are worrying Wall Street investors, as per the report by Bloomberg. Even top AI supporters admit the market is "frothy" but believe the technology will change the world -- curing diseases and transforming industries. However, no one really knows how AI will make steady profits yet. Many executives are spending heavily on AI out of fear of falling behind rivals -- not because profits are guaranteed. In January, OpenAI CEO Sam Altman announced a $500 billion "Stargate" AI infrastructure plan at the White House, shocking many with the price tag. Meta's Mark Zuckerberg followed by promising to invest hundreds of billions in AI data centers. Altman later said OpenAI could spend "trillions" on AI infrastructure, as stated by Bloomberg report. In September, Nvidia agreed to invest up to $100 billion in OpenAI's data centers.Analysts say this raised questions -- was Nvidia helping OpenAI mainly so it could keep selling its own expensive chips? Nvidia has invested in dozens of AI firms that often use the money to buy Nvidia chips -- a potential conflict of interest. The OpenAI deal was the largest of its kind so far. OpenAI may take on huge debt instead of relying fully on Microsoft or Oracle. Unlike those long-profitable companies, OpenAI expects to burn through $115 billion by 2029, according to The Information. Other tech giants are also borrowing heavily to fund AI infrastructure. Meta borrowed $26 billion to build a massive Louisiana data center nearly as big as Manhattan. JPMorgan and Mitsubishi UFJ are leading a $22 billion loan to Vantage Data Centers for a huge AI complex, as per the report by Bloomberg. Consultancy Bain & Co. said that by 2030, AI firms will need $2 trillion a year in revenue to cover computing costs. But their actual income could fall $800 billion short. Hedge fund manager David Einhorn warned, "There's a reasonable chance that a tremendous amount of capital destruction is going to come through this cycle." Many lesser-known firms are now joining the data-center boom. Nebius, a cloud firm from Amsterdam, signed a $19.4 billion deal with Microsoft. Nscale, a British firm once focused on crypto mining, is now working with Nvidia, OpenAI, and Microsoft on AI centers in Europe. Bloomberg's John Authers warned, "AI will have profound consequences... but there will be some pain ahead." Many investors still doubt the payoff from AI. An MIT study (August) found 95% of companies saw no return on AI investments. Harvard and Stanford researchers said employees are using AI to make "workslop" -- fake productive content that looks good but adds no real value. This "workslop" could cost large companies millions in lost productivity, as mentioned in the report by Bloomberg. AI companies like OpenAI and Anthropic believe more computing and bigger models will lead to Artificial General Intelligence (AGI) -- AI smarter than humans. But recently, they've seen diminishing results despite bigger models and spending. OpenAI's GPT-5, launched in August, got mixed reviews. Altman admitted, "We're still missing something quite important" to reach AGI. Chinese firms are flooding the market with cheaper AI models, putting price pressure on U.S. companies. That could make it harder for Silicon Valley to recover its massive AI investments. The AI data center explosion also threatens to strain national power grids due to huge electricity use. Sam Altman agrees there's a "bubble," but still believes AI is historic. He said, "Are investors overexcited? Yes. Is AI the most important thing to happen in a very long time? Also yes." Zuckerberg wrote in July, "Developing superintelligence is now in sight." Developers like OpenAI and Anthropic say they need much more computing power to meet growing demand, as per the report by Bloomberg. Anthropic said three-quarters of its clients use Claude to automate work. OpenAI launched a new tool, GDPval, to measure AI performance in real jobs. It said, "Today's best frontier models are already approaching the quality of work produced by industry experts." OpenAI CFO Sarah Friar said the firm could someday charge $2,000 per month for AI tools if they truly act like "a Ph.D.-level assistant." Zuckerberg said in September an AI bubble is "quite possible," but warned it's worse to not spend enough on AI. A market bubble is when prices rise too fast beyond real value -- followed by a crash. Bubbles go through 5 stages, displacement, boom, euphoria, profit-taking, and panic. They often burst when investors realize profits don't match the hype. In January, China's DeepSeek launched a cheap but powerful AI model. That caused a trillion-dollar selloff in tech stocks as investors panicked. Nvidia's stock fell 17% in one day. But soon, investors came back, and Nvidia's stock hit new highs -- over $4 trillion valuation by September, making it the most valuable company in the world, according to the report by Bloomberg. Just like the late 1990s, AI startups are raising massive money based on hype, not profit. In the dot-com era, telecom firms built too many fiber-optic networks, then crashed in 2001. Today, AI firms are building massive infrastructure and promising huge revenue that may not come. Venture capitalists are throwing money, luxury perks, and private jets at AI founders. Some startups are raising multiple mega-rounds in one year -- unsustainable for many, say reports. Bret Taylor, OpenAI chairman and CEO of Sierra, said, "It is both true that AI will transform the economy... and that we're also in a bubble, and a lot of people will lose a lot of money." He compared today's AI moment to the dot-com bubble, saying some firms will fail but others like Amazon or Google could emerge winners. Experts note differences -- today's big players (the Magnificent Seven) are rich and profitable, unlike 1990s startups. AI adoption is also faster -- ChatGPT already has 700 million weekly users. OpenAI's revenue may triple in 2025 to $12.7 billion, but it still won't be profitable until late in the decade. Despite never making a profit, OpenAI's valuation hit $500 billion, making it the most valuable unprofitable company in the world, as per the report by Bloomberg. Q1. Is there a risk of an AI market crash? Yes, experts warn that massive AI investments and hype could create a trillion-dollar bubble that may eventually burst. Q2. Why are tech companies spending so much on AI? Companies like OpenAI and Meta are investing billions in AI infrastructure to stay ahead, meet growing demand, and aim for long-term profits.

AI Gold Rush Is Fueling Corporate Debt Like Never Before | Investing.com UK

To add a little balance to my "The New Era Isn't a Bubble -- It's a Rerating of Reality," it's not that I'm naive to the risks. Every cycle hides its own fault line, and this one won't be any different. However, I think the tail risk will persist more deeply in the credit market. There's a strange echo in the market these days -- a low-frequency hum of capital that seems to come not from factories or consumers, but from the future itself. The AI boom, once sold as a profit flywheel of endless efficiency, has quietly become the world's most elaborate exercise in financial time travel -- where tomorrow's expected massive cash flow is used to fund today's infrastructure in the hope that belief alone will bridge the gap. At first, it looked self-sustaining: hyperscalers like Microsoft (NASDAQ:MSFT), Amazon (NASDAQ:AMZN), and Google (NASDAQ:GOOGL) reinvesting their monstrous free cash into data centers, GPUs, and AI training clusters. But something shifted. Oracle decided to break ranks, levering itself to the hilt -- a 500% debt-to-equity ratio -- to join the hyperscaler Olympics. Suddenly, what was once a cash-flow-fed growth story began to smell like a credit-fed arms race. It's as if a handful of companies are trying to build four nuclear plants' worth of power and calling it innovation. The irony? The "AI miracle" has generated barely enough revenue to pay for college essay subscriptions and a handful of enterprise pilots. Yet $500 billion in annual capital expenditure needs financing -- now. So the question that no one on CNBC seems willing to ask is: who's actually footing the bill for this moonshot? The answer, increasingly, is debt. What used to be called venture capital has been reborn as venture credit. Banks, private funds, sovereigns, insurers -- everyone is being enlisted to bankroll the AI gold rush. Morgan Stanley estimates nearly $3 trillion in global data center spend through 2028, half of which can't be funded internally. That leaves a $1.5 trillion hole. Enter private credit, collateralized data centers, and "AI infrastructure bonds" -- all dressed up in ESG-scented wrappers and sold to pension funds starved for yield. We've seen this playbook before: telecoms in 1999, shale oil in 2013, crypto mining in 2021. Each cycle begins with a story that promises exponential growth and ends with financiers discovering exponential math doesn't always apply to balance sheets. The new AI ecosystem -- a web of interlocked promises, circular contracts, and synthetic cash flows -- feels eerily similar. Oracle borrows to build data centers to host OpenAI workloads that don't yet exist, generating revenues that investors assume will. It's a hall of mirrors built on future belief -- and belief, like credit, is only as strong as its collateral. For the first time in modern market history, AI-related corporates now represent 14% of the U.S. investment-grade bond index, surpassing the banking sector. That's an extraordinary shift in the financial architecture of capitalism: tech has effectively become the new financial system. But here's the paradox -- these are not banks. They're not capitalized, regulated, or stress-tested like banks. They're industrial conglomerates disguised as innovation platforms, issuing debt at record-tight spreads (74bps on average) as if cash flow risk were a rounding error. Yet many are funding multi-decade energy and infrastructure projects with three-year paper. In trader terms, the street's running an unhedged carry trade on the future of AI. AI is a power story before it's a productivity story. Each new data center requires gigawatts of electricity and millions of gallons of cooling water. Tech/AI analysts peg the capital need at half a trillion dollars a year -- equivalent to rebuilding the global energy grid every decade. The irony is breathtaking: the technology meant to optimize the world's resources is consuming them at a pace that could destabilize credit markets. And like all leverage-driven manias, the narrative hides the fragility beneath. If even one leg of this scaffolding falters -- say, a breakthrough in quantum computing that renders current GPUs obsolete, or a China-built LLM that undercuts Western margins -- the equity bubble will deflate. But the real damage will come from the credit side: the loans, securitizations, and private notes written on top of revenue projections that never arrive. AI's great paradox is that it has become the world's most capital-intensive idea. We've turned "intelligence" into an industrial commodity, financed like shale or steel. The market is now pricing intelligence as if it were oil -- barrel by barrel, watt by watt -- ignoring that intelligence, unlike oil, doesn't always pay for its own extraction. And that's where the risk lies. The credit market has effectively securitized hope. Insurance funds and sovereigns are holding paper backed by projected AI revenues -- not realized profits -- in the same way subprime lenders once packaged NINJA loans as AAA tranches. Only this time, the collateral isn't a house; it's a GPU cluster in Nevada. We've entered the AI-as-debt era -- where intelligence itself is being leveraged, packaged, and sold as a financial product. This isn't to say the AI dream will implode. It may evolve. Technological breakthroughs could save it -- more efficient ASIC chips, new cooling systems, or genuine quantum computing leaps that cut costs. But the math is stubborn: unless productivity growth catches up, the sector is borrowing its way into an illusion of progress. When future cash is discounted back to the present at too low a rate, you don't get innovation -- you get asset inflation. Every generation's "new economy" starts with promises and ends with leverage. The question is no longer whether AI will change the world -- it's whether the world can afford the debt required to find out.

AI is becoming the 'magic fix' as America places 'one big bet' on it not being a bubble, market veteran warns | Fortune

A lot is riding on the AI boom, and it isn't just the stock market surge. AI is being touted as an elixir for a number of serious economic challenges, according to Ruchir Sharma, chair of Rockefeller International. In a Financial Times column on Sunday, the market veteran pointed out that the "immigration boom-bust cycle" that the U.S. is experiencing now is unprecedented in scale, swinging from a net gain of more than 3 million in 2023 to an expected trickle of just 400,000 this year. The drastic throttling in the labor force could slash U.S. growth potential by more than 20%. "Yet increasingly the response to this risk too is a shrug. AI is going to make human labor less necessary anyway," Sharma quipped. Meanwhile, the U.S. debt-to-GDP ratio is already at 100% and expected to continue galloping higher, topping the World War II-era record high in the coming years. But again, AI could come to the rescue by propelling economic growth enough to stabilize the debt. The global bond market even appears to be pricing in that scenario, Sharma said, pointing to surging yields for Japan, France and the U.K., even though they have smaller budget deficits than the U.S. does. "The main reason AI is regarded as a magic fix for so many different threats is that it is expected to deliver a significant boost to productivity growth, especially in the US," he added. In addition to the workforce and debt woes, AI could even ease inflation risks, including tariff-driven pressure, by enabling companies to raise wages but still keep prices steady, Sharma said. The hoped-for benefits of a productivity boom aren't totally far-fetched. The Congressional Budget Office estimated earlier this year that booting productivity growth by 0.5 percentage point each year for 30 years could make publicly held debt 113% of GDP by 2055, instead of 156%. And the U.S. has in fact enjoyed more productivity growth in recent years than other developed economies have, stoking further hype among investors that the lead will widen. America's AI narrative has helped global investors overcome the shock of President Donald Trump's trade war and "Liberation Day" tariffs, which triggered a sudden exodus out of U.S. markets. But the money quickly came back, and Sharma said foreigners plowed $290 billion into U.S. stocks in the second quarter and now own 30% of the market. "In a way, then, America has become one big bet on AI," he said. Excluding AI-related stocks, European markets have actually been beating the U.S. this decade, and the outperformance is spreading to other sectors. "What that suggests is that AI better deliver for the US, or its economy and markets will lose the one leg they are now standing on," Sharma warned. He's not the only voice sounding the alarm. Lisa Shalett, Chief Investment Officer for Morgan Stanley Wealth Management, wrote on September 29 that "it's hard not to still see ... a boom driven by a one-note narrative." Since ChatGPT's launch, Shalett noted, what she considers "AI data center-ecosystem stocks" have accounted for roughly 75% of S&P 500 returns, 80% of earnings growth and 90% of capex growth. "It's difficult to ignore the market's reliance on AI capex," she concluded. For now, Wall Street seems happy to ride the wave. On Monday, OpenAI's announcement that it's taking a stake in chipmaker AMD sparked another stock market rally. Analysts are also hiking price targets for other hot AI plays like Nvidia as well as the overall S&P 500. And while the recent string of record highs has fueled concerns about a bubble, certain metrics indicate that the AI boom isn't yet at dotcom-bust levels. Others still see conditions getting frothier. Evercore ISI analyst Julian Emanuel said in a note on Monday that he now sees 30% odds of the S&P 500 soaring to 9,000 at the end of next year in a "bubble scenario," up from 25% odds just a few weeks ago. His base case is for the index to reach 7,750 by then, representing a gain of 15% from currently levels.

Analysts warn AI hype is a 'Red Flag' -- bubble now bigger than 2008 Subprime crisis

AI bubble 2025: Analysts are raising serious concerns that the artificial intelligence boom has become one of the biggest financial bubbles in modern history. The MacroStrategy Partnership, an independent research firm, found that the AI bubble is now 17 times larger than the dot-com bubble of the 1990s and four times bigger than the global real-estate bubble that caused the 2008 crash, as per a report. The note was written by Julien Garran, who previously led the commodities strategy team at UBS. Garran and his team argue that companies have overhyped AI's true capabilities, as per a Common Dreams report. He pointed to data showing that adoption of large language models (LLMs) by major businesses has already begun to decline. Garran also said that ChatGPT may have "hit a wall," noting that its latest version costs ten times more but doesn't perform noticeably better than earlier releases, according to the Common Dreams report. ALSO READ: Yes, you are feeling it: 22 states are officially in a recession -- is yours on the list? He warned that the economic fallout could be severe. Garran wrote in the report, "The danger is not only that this pushes us into a zone 4 deflationary bust on our investment clock, but that it also makes it hard for the Fed and the Trump administration to stimulate the economy out of it," as quoted by Common Dreams. Dario Perkins, managing director of global macro at TS Lombard, shared similar worries in an interview with Axios. He said tech companies are taking on massive debts to build AI data centers, comparing the trend to what happened during the dot-com and subprime mortgage bubbles, as per the Common Dreams report. He pointed out that the big tech companies, don't care whether the investment has any return, because they're in a race," and added that, "Surely that in itself is a red flag," as quoted in the report. ALSO READ: Behind on retirement? Smart ways to catch up fast before it's too late Goldman Sachs CEO David Solomon also cautioned. Speaking at the Italian Tech Week conference, he said he expects a "drawdown" in the stock market within the next couple of years because of the huge sums being funneled into AI projects, according to the Common Dreams report. He said, "I think that there will be a lot of capital that's deployed that will turn out to not deliver returns, and when that happens, people won't feel good," as quoted in the report. While Solomon did not definitively declare AI to be a bubble, but he did say some investors are "out on the risk curve because they're excited," which is a sign of a financial bubble, as reported by Common Dreams. Amazon CEO Jeff Bezos, also at the conference, said that there is a bubble in the AI industry. However, he added that AI will still bring major benefits to humanity. Bezos said, "Investors have a hard time in the middle of this excitement, distinguishing between the good ideas and the bad ideas," pointing at the AI industry, adding, "And that's also probably happening today," as quoted by Common Dreams. Perkins did not predict when the AI bubble might burst but said it's nearing its peak. He said, "I wouldn't touch this stuff now," and added that, "We're much closer to 2000 than 1995," as quoted in the report. What are experts saying about the AI industry? Experts warn the AI boom has grown into a massive financial bubble that could trigger a global downturn, as per the Common Dreams report. How big is the AI bubble compared to past crises? The AI bubble is reportedly 17 times larger than the dot-com bubble and four times bigger than the 2008 crash.

Why the AI Boom May Defy History: 4 Reasons This Time Could Be Different | Investing.com UK

For months, investors have asked a familiar question: Is artificial intelligence experiencing a bubble destined for the same catastrophic fate as the dot-com crash of 2000? With hyperscaler capital expenditure surging past $320 billion annually and AI stocks commanding valuations that eclipse even the TMT (Telecom, Media, and Technology) bubble peaks, the parallels seem undeniable. Yet beneath the surface, critical differences suggest today's AI surge may follow a fundamentally different trajectory than its predecessors. In the late 1990s, many internet business models depended on infrastructure that didn't exist yet. Dial‑up connections were slow, cloud computing was nascent, and the digital "plumbing" couldn't carry the weight of ambitious plans. Today's AI build‑out sits on top of mature cloud platforms, specialized chips, and global high‑bandwidth networks. Crucially, AI capabilities are visible and deployable now, from copilots in software development to quality control in manufacturing. These are production use cases that move beyond hypotheticals. That maturity lowers execution risk. The difference shows up in throughput and tooling. Hyperscaler data centers now run specialized accelerators at a global scale, with low‑latency interconnects and production‑grade MLOps. Companies can prototype, iterate, and ship AI features within standard release cycles, not multi‑year bets on unproven infrastructure. That enables more measured, ROI‑focused adoption than the build‑it‑and‑hope model common in the late 1990s. One hallmark of the dot‑com era was the focus on vanity metrics over earnings. In this cycle, enterprises are deploying AI to cut costs and speed up work. Studies and early rollouts point to double‑digit productivity gains for developers, with a meaningful share of production code now assisted by AI. Case studies from large companies, including logistics optimization in retail, defect detection in autos, and document processing in finance, show measurable outcomes. Spending patterns reflect that shift. Enterprise AI investment has been expanding steadily, and independent estimates from industry and consulting firms suggest large potential economic value as adoption broadens. While those forecasts will be debated, the direction of travel toward demonstrable ROI is clear. Early data points help frame the scale. Some enterprise studies have reported roughly 31.8% productivity gains for developers using AI copilots, with user satisfaction rates above 80%. Among top adopters, code volume increased materially, and AI‑assisted code now accounts for a significant share of output. Outside software, operations teams report concrete wins: retailers citing logistics savings, automakers measuring lower defect rates, and banks accelerating document review, incremental improvements that add up at scale. Aggregate spending is as follows. Industry trackers peg enterprise AI outlays in the hundreds of billions of dollars in 2024, with forecasts pointing higher as pilots move to production. Broader analyses estimate a multi‑trillion‑dollar annual value creation potential once adoption diffuses across functions. The range is wide, and outcomes remain uncertain, but results align with early case studies. AI is getting cheaper to run and build. Training costs for competitive models have fallen sharply, and new architectures are lowering inference costs. In many markets, falling unit costs reduce revenues. In AI, lower prices have coincided with faster growth in usage. This pattern aligns with the Jevons Paradox, a concept economists observed in energy markets during the Industrial Revolution. Recent industry tallies show AI server and storage revenues hitting record highs while per‑unit costs decline. Lower barriers to experimentation have pulled more developers and companies into the market, expanding the total pie. The response from the largest platforms supports this point. After efficiency gains, major cloud providers raised their AI infrastructure budgets, signaling expectations of sustained demand. Overall, spending on computing has grown. Industry reports tallied AI server and storage revenues of around $244 billion in 2024, up from roughly $89 billion in 2023, a period that coincided with lower unit costs. Enterprise AI spending is projected to continue rising, and developer adoption surveys show a majority of organizations now experimenting with AI tools, up from a minority just two years earlier. Efficiency appears to expand the addressable market. The Jevons Paradox in AI: As inference costs dropped 90% from 2022 to 2025, total AI spending surged over 400%. Who is doing the spending matters, and so does the motivation. Today's leading investors in AI are mega‑cap tech companies with strong cash flows and diversified businesses. That reduces financing risk compared with debt‑fueled booms. Policy also matters. The U.S. and China view AI leadership as a strategic priority, which supports continued investment through market cycles. While that doesn't eliminate commercial risk for individual companies, it does create a wider floor for the ecosystem than earlier tech booms. Budget intentions reflect that strategic lens. Following notable efficiency gains, large platforms outlined increased infrastructure plans for 2025, with combined guidance well above $300 billion across cloud and AI build‑outs. The investments are generally equity‑funded from substantial cash flows, which lowers systemic risk. Market commentators note that even missteps by individual firms would likely be absorbed within diversified businesses, reducing the odds of a sector‑wide crunch. Hyperscaler AI Infrastructure Spending Accelerates: 2024 vs 2025 capital expenditure (Capex) across the four major hyperscalers (Microsoft, Amazon, Google, Meta). Total spending is increasing from $202B to $270B, a 34% year-over-year jump. None of this guarantees a smooth path. History suggests periods of overbuild, pricing pressure, and shakeouts as competition intensifies. Key risks include: Additional watch‑items: supply chains for advanced chips; power and cooling constraints as data center density rises; talent bottlenecks in AI safety, reliability, and tooling; and the possibility of commoditization in undifferentiated application layers. Competition between closed and open‑source models could pressure margins, and geopolitics, from export controls to data‑sovereignty rules, may alter the cost curve and go‑to‑market plans. Many investors aim to balance opportunity and risk. Strategies include: Dot-Com Bubble vs AI Boom: 32% productivity gains vs 5%, 85% revenue-positive companies vs 35%, 9/10 infrastructure readiness vs 3/10, 78% measurable ROI vs 15%, and 90% equity funding vs 45%. Several catalysts could shape the next leg of the cycle: The AI boom shares surface similarities with past bubbles, including fast‑rising valuations and intense attention, while the underlying conditions differ in important ways: mature infrastructure, visible productivity gains, efficiency that expands demand, and institutional support. Corrections are possible, and weaker players are likely to be winnowed out. The foundations suggest today's cycle may endure longer and prove more consequential than a simple bubble analogy implies.

The AI bubble will pop. Intelligence won't | Fortune

Oracle's 43% stock surge in a single day should make investors uneasy. This isn't a meme stock or speculative startup, it's one of America's largest tech firms, suddenly trading at bubble-era valuations. The AI boom has inflated the S&P 500 and Nasdaq to record highs. This run-up is the latest development that has led investors to wonder: Will the AI bubble pop? But what won't pop is intelligence itself. While Wall Street bids up mega-models with billion-dollar burn rates, AI is producing measurable returns elsewhere in transformational, if less glamorous, ways. Take Austin, Texas, where an on-premise AI-system helped local government process building permits in days instead of months. No spectacle. No headlines. Just efficiency gains that will outlast the market cycle. That's the point too often missed in the frenzy. Mega-models attract headlines, consume billions in capital, and struggle to demonstrate sustainable economics. Meanwhile, smaller, domain-specific systems are already delivering efficiency gains, cost savings and productivity improvements. The smart play isn't to abandon AI, but to pivot toward models and deployments that will endure. We've seen this movie before. Netscape once symbolized the internet revolution. Its spectacular IPO made headlines; its decline made history. But the collapse of early web darlings didn't kill the internet, it revealed that the real value lay not in browsers but in the infrastructure beneath them. AI stands at the same crossroads today. The platforms consumers know best -- ChatGPT, Gemini, Claude -- are extraordinary feats of engineering, but they don't represent sustainable AI. They are breathtakingly expensive to run, but free or cheap to use. They deliver entertainment and convenience more than enterprise value. It's fun to have ChatGPT turn out a poem, nice when it helps you smooth out an email -- but it's not mission-critical. Economically, the hyperscale model doesn't hold. Training and maintaining ever-larger systems yields diminishing returns while costs escalate into the billions. That's why GPT-5 landed with a shrug. Scale alone is no longer impressive. So what is? The answer lies in focused deployments. Austin's permitting office achieved in weeks what bureaucracy had delayed for years. Healthcare systems are running diagnostic models fine-tuned for their specialties that outperform general-purpose LLMs. Financial firms are already relying on BloombergGPT, trained on market data, which delivers better results in its domain than larger consumer platforms. These applications generate tangible ROI and do so sustainably. The principle is simple: a massive general-purpose model can do many things at a passable level, but it rarely excels. A leaner system, built for a specific function and deployed thoughtfully, can deliver speed and accuracy where it matters most and at a fraction of the cost. This is the strategic, cost-effective way forward: integrate AI in tactical ways that directly serve the business, rather than chasing the illusion of a one-stop shiny new AI technology. Think of it as staffing a project: 100 average consultants won't outperform five experts. Where the data lives matters just as much. Lighter models can be optimized to run locally on edge devices or inside secure enterprise facilities instead of depending on costly, centralized infrastructure. At webAI, for example, we've been able to shrink models by nearly a third while preserving accuracy. That changes the economics completely. Instead of routing every query through an expensive cloud data center, intelligence sits closer to the data it serves, making it cheaper, faster, more resilient and more secure. Just as important, enterprises retain ownership of their data and the insights built on it, which is not possible when relying solely on hyperscale providers. Companies tied exclusively to mega-models are exposed to spiraling costs, energy scrutiny, and security vulnerabilities. Decentralized, specialized AI avoids those traps. IT also offers resilience and puts businesses on firmer ground for the regulatory scrutiny that is sure to come. With this in mind, smart techno-optimists and AI investors need not panic when headlines warn of an "AI winter." Yes, some companies will collapse under the weight of unsustainable economics, just as many did after the dot-com crash. But AI itself isn't going away. It's evolving toward networks of specialized systems that work together more like a city grid rather than a skyscraper. For executives, the takeaway is clear: avoid chasing scale for its own sake. Instead, invest in AI systems that are efficient, close to your data, and tailored to specific business needs. Build for sustainability, not spectacle. When the next AI earnings cycle sends markets into another frenzy, remember Austin's building permits. The businesses building lean, domain-specific intelligence won't be watching their valuations with the same anxiety. AI isn't going back in the box. But the future won't be bigger at all costs -- it will be smarter, leaner and built for staying power.

AI startup valuations raise bubble fears as funding surges

SINGAPORE (Reuters) -Artificial intelligence startups are attracting record sums of venture capital, but some of the world's largest investors warned that early-stage valuations are starting to look frothy, senior investment executives said on Friday. "There's a little bit of a hype bubble going on in the early-stage venture space," said Bryan Yeo, group chief investment officer at Singapore sovereign wealth fund GIC, as part of a panel discussion at the Milken Institute Asia Summit 2025 in Singapore. "Any company startup with an AI label will be valued right up there at huge multiples of whatever the small revenue (is)," he said. "That might be fair for some companies and probably not for others." In the first quarter of 2025, AI startups raised $73.1 billion globally, accounting for 57.9% of all venture capital funding, according to PitchBook. The surge was driven by funding rounds like OpenAI's $40 billion capital raising, as investors raced to catch the AI wave. "Market expectations could be way ahead of what the technology could deliver," Yeo said. "We're seeing a major AI capex boom today. It is masking some of the potential weaknesses that might be going on in the economy." Todd Sisitsky, president of alternative asset manager TPG, said the fear of missing out is dangerous for investors, though he added that views were divided on whether the AI sector had formed a bubble. Some AI firms are hitting $100 million in revenue within months, he said, while others in early-stage ventures command valuations at between $400 million and $1.2 billion per employee. He said that was "breathtaking." (Reporting by Yantoultra Ngui; Editing by Thomas Derpinghaus.)

Spending on AI is increasingly fuelled by debt and will be marginally 'negative' for corporate credit quality, Goldman Sachs says | Fortune

The S&P 500 rose to a new all-time high yesterday, up slightly to hit 6.715. AI stocks led the way, again, with Palantir going up 1.1%, Nvidia rising 0.91% and Amazon climbing by 0.81%. In the absence of macro jobs data due to the U.S. government shutdown, traders are piling into AI stocks, ING told clients this morning. "Financial market volatility is falling across the board, partly driven by the US government shutdown and the delay to key data releases such as the September jobs data. Instead, investors remain transfixed by the AI-driven rally in megacap tech shares, which shows no signs of slowing." S&P futures were up 0.12% this morning, premarket. Stocks are benefiting from the massive amount of capital expenditure (capex) coming from companies that are investing heavily in AI. "For now, AI is being driven by a handful of US Big Tech players spending almost $350 billion on Cap-Ex this year with now the cavalry coming as more enterprises and governments from around the world get into the AI spending game. The Middle East is a perfect example as now Saudi and UAE are diving into the deep end of the pool with Nvidia and US hyperscalers set to build out massive AI driven data centers to fuel the use cases for the coming years," Wedbush's Dan Ives told clients in a note this morning. But where is all that money coming from, exactly? Goldman Sachs notes that, increasingly, it is coming from debt. Over the last three years, most capex came from the cash sitting on tech companies' balance sheets. But now those companies are increasingly borrowing money to fun AI build-out. "These AI-related issuers have accounted for $141 billion in corporate credit issuance in 2025 to-date, eclipsing full-year 2024 gross supply of $127 billion," Lotfi Karoui and his colleagues wrote in a note seen by Fortune. The shift will marginally reduce the quality of corporate credit, Karoui wrote. "The read-through for credit markets is, on the margin, negative, in our view. While not yet a cause for alarm, given both the high cashflow generation and low leverage among large tech companies, the shifting funding mix of capex beyond cash is worth monitoring." The note is interesting because there has been a lot of talk on Wall Street recently about the quality of corporate debt. Two auto-industry companies, First Brands and Tricolor, recently collapsed into bankruptcy because they issued too much debt to lenders who thought they were safe. Jim Chanos and PIMCO President Christian Stracke have both recetly bemoaned cracks in the private credit market.