Tech Giants Slam the Brakes on AI Spending as Tokenmaxxing Drains Budgets Without Results

Amazon Is the Latest Tech Giant to Face the Consequences of AI 'Tokenmaxxing'

Big companies are pulling back from using AI for anything and everything as costs go up. Amazon became the latest tech giant to admit that maybe all AI all the time is not the best idea. The retail giant has reportedly shut down its employee-led AI leaderboard (known as Kirorank), an internal mechanism the company used to encourage employees to use AI more often. According to a report by the Financial Times, Amazon's decision was based on two factors. The first was cost. More companies are using more AI; the additional usage, measured in tokens, is causing prices to skyrocket. The second reason is due to employees engaging in a practice known as tokenmaxxing, where they make AI do menial tasks to increase token usage so that they score better on the leaderboard. In short, Amazon was spending a lot of money on AI that wasn't really doing anything. These two problems reportedly cost Amazon enough money that the company decided to scale back its AI usage by removing the leaderboard and calling for a stop to tokenmaxxing. "One of the internal dashboards, called Kirorank, was recently created by a group of employees who wanted to drive awareness for how AI can accelerate work, and was never intended to promote the use of AI for usage's sake," an Amazon spokesperson told CNET in an email. "The beta dashboard was not a formal or approved tool, and has since been deprecated. We're focused on AI adoption and sharing best practices to celebrate innovation and operational efficiency gains across the company, and we're proud of the way our teams are embracing this technology." Soothsayers probably saw this one coming as Amazon's battle with tokenmaxxing was made public earlier in May. A leaked employee memo from Dave Treadwell, Amazon senior vice president, asked employees to stop "using AI just for the sake of using AI" in response to employees overusing AI to get to the top of the leaderboards. Amazon says it measured token usage to understand cost and efficiency but discouraged using those metrics to measure developer productivity. Companies are pulling back on AI overuse Amazon is one of several companies to admit that maybe they're using a little too much AI. Meta forcibly closed down an employee-run AI leaderboard in April after employees engaged in tokenmaxxing to compete for the "Token Legend" status. In a Rapid Response interview last week, Uber Chief Operating Officer Andrew Macdonald admitted that the rideshare company was struggling to justify further AI costs after a viral interview from Uber Chief Technology Officer Praveen Neppalli Naga revealed that the company had blown through its entire 2026 AI budget in just a quarter of a year. It's difficult not to label these decisions as a trend. Microsoft began canceling Claude Code licenses in early May. According to the Wall Street Journal, Salesforce, DoorDash and several other big names have also gone from throwing AI at everything to rationing it amid soaring costs with lackluster returns. Despite the pullback, generative AI use remains at an all-time high. Google announced that Gemini had jumped from 480 trillion tokens per month in May 2025 to 3.2 quadrillion tokens per month as of May 2026. One cause is that agentic AI, coding tools and always-on tools like OpenClaw -- all of which have surged in popularity this year -- burn through far more tokens than basic text prompts and responses from chatbots. "I would say [companies pulling back on AI usage] isn't surprising, but probably not enough of a slowdown that it is going to burst the generative AI bubble that we seem to be in," said Jackie Rees Ulmer, dean of the Ohio University College of Business, in an email. "As companies get better at sorting the applications that provide real value versus using AI just for the sake of using AI, demand will only increase." This seems to be the rallying cry of these businesses. Will McGough, chief investment officer at Prime Capital Financial, told the Wall Street Journal the same thing, stating that companies were still very much "figuring things out" when it comes to effective AI use. Ulmer says that the path forward is education, and that she encourages her students to learn more about AI applications relevant to their fields, but to always "double down" on "human skills, such as critical thinking and communication."

Uber president says AI spending is getting 'harder to justify'

After reportedly exhausting its annual AI budget just four months into 2026, Uber is now questioning whether it's actually seeing meaningful returns on its investments. In an interview with Rapid Response, Uber president and chief operating officer Andrew Macdonald said the company isn't seeing a connection between rising token consumption for Claude Code and more useful features being delivered to consumers. "That link is not there yet, right? I think maybe implicitly there is more that is getting shipped, but it's very hard to draw a line between one of those stats and, 'Okay, now we're actually producing 25 percent more useful consumer features,'" said Macdonald. "I think over the coming quarters and years, maybe that will become clearer, but I think today it's hard, even if some of the underlying metrics are trending in a really astronomical direction." Uber spent $3.4 billion on research and development efforts in 2025, 9 percent more than it had spent the previous year. Earlier this month, Uber CEO Dara Khosrowshahi said the company was making up for its increasing AI investments by hiring fewer human employees. "We're going to have to start talking about token consumption and the associated cost versus headcount," said Macdonald. "So if you're not actually able to draw a direct line to how much useful features and functionality you're shipping to your users, that trade becomes harder to justify."

AI costs begin to bite as agents may increase token demand by 24 times, says Goldman Sachs report -- Uber and Microsoft among companies feeling the bite of tokenized billing

Microsoft and Uber are both considering refining their AI strategies as costs mount. Major tech companies are struggling to justify the skyrocketing prices of heavy AI usage, with even major tech firms like Microsoft and Uber looking at changes to their AI process. Following the recent viral post from Uber CTO Praveen Neppalli Naga that the company had blown through its entire 2026 AI budget in just a few months, Uber's Operations chief, Andrew Macdonald, said that token usage just didn't seem to have a direct correlation with useful consumer features. Microsoft began revoking its developers' access to the Claude Code programming assistant earlier this month, with plans to move them over to the internal Copilot CLI tool by June 30. Although that has been framed as consolidating its teams onto the tools it's developing, it also comes right at the end of Microsoft's fiscal year, suggesting it may have also been a move to cut costs before the new year. Worsening matters, Goldman Sachs estimates that Agentic AI could see token use increase by over 24 times in just the next few years. There appears to be a growing disconnect between AI needs, AI wants, and the reality of what AI companies can actually afford as costs mount. Tokens and trade offers We've been hearing reports for months about how companies and CEOs are struggling to find the tangible benefit of heavy AI deployment. Uber appears to be the latest AI boosting company to have this come to Jesus moment, following the CTO's explosive claims of annual budgets being wiped out in mere months. In the interview with Business Insider, Andrew Macdonald lamented that there just wasn't a clear correlation between the money Uber was investing in AI use and real consumer feature development. Having talked to the senior engineers, he said there was no link between higher token usage and a proportional increase in consumer features with real benefits for their customers. Although he admitted more code was being shipped, it "was very hard to draw a line" between that and improvements in the software. Meanwhile, after opening up its workers to Claude Code subscriptions in December last year, Microsoft is now clawing that back in what's seen by many as a financial move, as much as a consolidation. Microsoft also recently announced the switch of Copilot on GitHub to token-based billing, as the cost of running the tool ballooned earlier this year. A major reason for this is the explosive growth in agentic AI use. These agents can eat up more than 1,000 times the tokens of a single AI chatbot. Are more tokens really the answer? Nvidia CEO Jensen Huang famously said in March this year that if an Nvidia engineer on $500,000 a year wasn't using at least $250,000 of tokens in that same period, he'd be alarmed. This isn't a rare sentiment, though. Many company CEOs are now bragging about the extent of their AI use, as if that alone equates to performance increases. As Business Insider reports, Airbnb's CEO proudly told investors that 60% of the company's code was now AI-generated. Chime claimed it was shipping 84% AI code earlier this year, and even Google is claiming 50% of its code is AI-generated (though crucially, always checked by a human engineer). Yet these numbers sound very similar to those of Uber. In the CTO's shocking report of budget runaway, they claimed over 80% of Uber software engineers were using agentic AI, and over 60% of the code was AI-generated. Even then, it's not worth the cost. And those costs can be extreme if the guardrails are removed. OpenClaw creator and now OpenAI employee, Peter Steinberger, recently announced his team of three people had spent over $1.3 million in tokens in a single month running a suite of agentic AI tools. This very much reinforces the idea that the cost of AI is rising above that of the workers it's supposed to be replacing. That makes many of the layoffs laid at the feet of AI efficiency and productivity increasingly shaky, unless these companies are simply racing to the bottom. Or at least racing to new hardware. Goldman Sachs' recent AI agent report suggests that the massive efficiency gains coming from next-generation inferencing chips would make AI use so much cheaper that investment can continue unabated, and profit should follow, with AI agents increasing the revenue at AI companies enormously. Faster, more efficient hardware will take too long Nvidia will talk up its Vera Rubin platform at Computex and will officially launch it later this year. It improves AI performance by several times over, uses a new process node, and will reportedly offer as much as 10 times the performance per watt, making it dramatically more efficient than its predecessors. Such huge gains would give the AI companies that first deploy these cards an enormous advantage over the companies still running Blackwell hardware, and even more so over older Hopper designs. But over 50% of the data center projects announced with Blackwell hardware in mind have been cancelled or delayed, and of those that do complete in the next year, just how keen are the developers going to be to replace those GPUs after they've barely gotten started? In late 2025, Google, Oracle, and Microsoft all adjusted their plans for hardware in the other direction entirely, suggesting they would make it run for six years before replacing it. That seems impossible to square away with ambitious AI plans and hardware leaps every year. More tokens on less efficient hardware The reality is, even as some token costs are falling, the explosion in the number of agentic AI requires cannot be offset by hardware efficiency gains that are many years away from reaching effective deployment, if they ever get to the scale needed to catch up with this ramp-up in AI demand. That means in the short term, even major companies like Microsoft and Uber are restructuring their use of AI to figure out how to continue using it at scale without nuking their budgets in the process. If those companies can't figure out how to afford it, it's increasingly difficult to imagine how the rest of us will be able to. And if usage drops because of rising costs, the AI companies are never going to find the short-term profit they need to offset the enormous infrastructure spending they're still trying to justify.

Uber chief warns no link yet between AI tokenmaxxing and shipping successful products -- company pumps the brakes on all-out AI spending

Uber CTO previously noted that the Claude Code budget for the entirety of 2026 had been spent by April. Uber President and COO Andrew Macdonald has warned that there is not yet a link between higher AI token usage and an increase in useful consumer features, seemingly pumping the brakes on 'AI tokenmaxxing.' Speaking on the Rapid Response Podcast (via Business Insider), Macdonald said, "That link is not there yet, right?" when it comes to using AI to ship features useful to consumers. Macdonald's headlining quote was taken from a part of the podcast where the discussion concerned Uber working to shape products with an eye on "what's better for the consumer." The conversation points to LLMs being used to try to hit the key marketing goal of addressing consumer wants and needs better than any competitor. The Uber COO noted that "We're working with pretty much all of the large model companies." However, the issue is that management isn't seeing a clear link between spending on AI services and successful products shipping. This may simply be because, so far, "there hasn't really been anything that's taken off yet." The interviewer and host, Bob Safian, highlighted the disappointment voiced by Duolingo employees and management regarding using AI in the workplace. Employees voiced concerns that AI was being pushed for the sake of it, while it introduced new workloads such as checking and reinforcing tasks. Duolingo management heard them and now understands that AI/LLMs don't fit everywhere. Uber's Macdonald nodded along and interjected that "the headline stats make your head explode" when companies discuss AI usage. But he also cautiously indicated that management should ask what productivity gains were delivered and what new products were AI-driven. So, Uber management isn't against using LLMs from the top providers, and it sounds like they will continue to do so. But there may be a reckoning for this technology if a clear link between spending on it and performance doesn't emerge. News of Uber's spending on AI/LLMs went viral last month. Uber CTO Praveen Neppalli Naga told The Information that his company had already blown through its Claude Code budget for 2026 by April. That incident likely caused a few heated discussions in the Uber boardroom. Perhaps Macdonald's interview provides a window into a philosophy change within Uber and one possible alternative to tokenmaxxing. Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds.

A company spent $500 million in one month after forgetting to set AI usage limits

Additionally, we're starting to see pushback from corporations and consumers about rising AI costs. It's been an unexpected shift in opinions on AI, with corporates recently pushing back on its use due to unsustained output despite mounting API costs. Leaders at brands such as Costco, Delta Airlines, and IBM have recently echoed their concerns about AI and a preference to retain the human workforce, especially as others, such as Amazon, Meta, and Microsoft, continue to cut jobs. Most recently, comments from Uber's new COO, Andrew Macdonald, about AI-related costs and token usage not improving workers' productivity as they should were heard, and mostly appreciated, across the internet. This was followed by reports that Uber engineers had already exhausted their AI budget for 2026. Turns out Uber may not be the only company struggling to keep its AI budget in check. According to an Axios report (paywalled), an unspecified company burned through roughly $500 million in Claude credits after failing to put guardrails on usage. This, among other incidents, is starting to push corporate leaders to evaluate whether AI is truly delivering the value they first assumed. The report also notes that the corporate cadre is starting to ditch "tokenmaxxing," a term used to describe the tendency to burn through AI credits as fast as possible. To counter that sentiment, AI biggies, including Google, have been building models and inference techniques that are more cost-efficient. Adding to this, a recent Gartner report says that inference costs for generative AI models in 2030 will be only a tenth of what they were in 2025. However, it's important to note that usage may also grow exponentially, especially as our reliance on AI agents increases and processes become more complex. The report also predicts token usage to expand anywhere from 5 to 30 times the current usage. Providers, including Google and Anthropic, have recently also shifted to a usage-based billing and stricter usage limits, which has rightly caused agitation among non-enterprise users. Even companies betting their future on AI, such as Microsoft, are getting away from their tokenmaxxing approach. Earlier this month, Microsoft reportedly began canceling Claude subscriptions and discouraging employees from using it too much, just six months after it began pushing various workers across different profiles to vibe-code more. This is yet another proof that overreliance on AI is costing enterprises more than the benefit it offers. It's difficult to predict whether we could see a complete reversal of the initial AI fervor. To be fair, it's highly unlikely as well. But we could certainly see companies budgeting their AI usage and restricting it to certain activities rather than giving employees a free rein. The AI bubble, so to speak, may not burst, but the AI dream could be beginning to end already.

AI Investment Is 'Harder to Justify' as Productivity Returns Lag, Uber COO Says

Uber does not see a link between productivity gains and its colossal AI spending commitments, according to the company's chief operating officer. Like much of the rest of the tech industry, Uber went all in on AI this past year. Company executives had employees across divisions embed AI into their workflows and reported that they were relying on AI agents for 10% of all code changes. While Uber doesn't outwardly subscribe to this practice, many tech giants have reportedly begun incentivizing more AI use among their employees with internal token use leaderboards. In March, Nvidia CEO Jensen Huang argued that if you hire a software engineer for $500,000 a year, they should be consuming "at least $250,000 worth of tokens" in that same time frame. But tokens cost money, and it's supposed to come from somewhere. Most executives have found that "somewhere" to be their own employees. Last month, Uber CTO, Praveen Neppalli Naga, told The Information that Uber had already exceeded its 2026 AI budget within the first four months of the year. Shortly after, at the company's earnings call earlier this month, executives said that Uber would be upping that AI spending even further and slowing down hiring to help pay for it. But now, Uber COO Andrew Macdonald says that those towering financial commitments are not directly correlating with real, tangible productivity gains, and it's making it tougher to justify the headcount reductions. "You talk to your senior engineering leaders, and you're saying, 'Okay, how many projects that were on the cutting room floor got moved above the line because of the productivity gains, because 25% of our code commits were via Claude Code last quarter,' that link is not there yet," Macdonald told the Rapid Response podcast over the weekend. "If you're not actually able to draw a direct line to how much useful features and functionality you're shipping to your users, that trade becomes harder to justify." Macdonald's comments echo long-held worries about the AI boom and its murky productivity returns. Companies across industries are spending staggering amounts of money to adopt AI and have it automate tasks across workflows, hoping for unbelievable returns on productivity and new highs for profit. The thinking among executives goes that because AI can work super fast and around the clock, and it doesn't have the same needs as a human employee does, then surely it must be the wiser economic decision to have agents take over workflows instead. But while AI can feel like it's free for those who are using it, Macdonald said, "somebody's paying the bill." The AI industry has promised that the bill and the sacrifices made to pay for it will be worth it, but that scenario hasn't panned out as quickly as promised. Another way the Uber executive says he's seen AI's lack of return is through the rise of commerce AI agents. As agentic AI gains ground, some experts started sounding the death knell for app providers like Uber and DoorDash, whose business models they say will be completely disrupted by agentic AI assistants that will shop on behalf of the average user. "We're working with pretty much all of the large model companies as they roll out, try to roll out commerce, and there hasn't really been anything that's taken off yet," Macdonald said. "It doesn't mean it won't happen, but a year ago at our board meeting, we were worried that by this point, a year later, we'd be totally disaggregated, because all commerce is going to be flowing through the large model codes in the form of chatbots, and that just hasn't played out yet." Uber is only one, though rather major, addition to a growing trend that's reportedly forming: even though the entire business world seems bullish on AI, its productivity returns are not following suit for everyone. If that trend continues, that has potentially disastrous consequences for the unprecedented AI infrastructure buildout this country is undertaking and the entire U.S. economy that AI spending has been propping up.

Corporate America enters its AI reckoning

Why it matters: Companies that rushed to embrace AI are now confronting ballooning IT costs, uncertain productivity gains and growing employee skepticism. Driving the news: Microsoft canceled most of its Claude Code licenses, in part over costs, according to The Verge, and Uber's COO said AI costs are getting "harder to justify." * An AI consultant tells Axios one of their clients recently spent half a billion dollars in a single month after failing to put usage limits on Claude licenses for employees. * Companies are citing AI's ability to automate jobs as a cause for layoffs, though Anuj Kapur, CEO of CloudBees, told Axios that workforce cuts may simply be "the only lever they can pull" to offset their AI bills. * Consumer sentiment around AI is also nosediving, and employees are rebelling against the use of the technology at work. What they're saying: The enterprise is undergoing a "healthy swing" away from AI overuse -- or "tokenmaxxing," the push to burn as many AI tokens as possible -- Ali Ansari, CEO of model training firm Micro1, told Axios. * Ansari hopes this correction will push companies toward more efficient AI use. * While the market views these tools as working equally well across the enterprise, Ansari says "the reality of AI right now is that it only works for coding." * That disconnect can drive up IT bills without leading to high return on investment in agents, he said. Friction point: Corporate AI adoption is running into four unique problems. * Use cases: "Most people default to automating tasks they dislike rather than tasks most valuable to the company," Sophia Velastegui, CEO of Velastegui Ventures and former chief AI officer at Microsoft told Axios. Instead, they should focus on using AI to drive revenue. * Costs: One CTO told Axios that employees were using AI models to check the weather. That gets expensive fast: Enterprise AI plans are not truly 'all you can eat,' and even simple chatbot queries can carry heavy token costs. * Humans: We are the bottleneck to more efficient adoption, as we're still catching up on AI. Leadership isn't always helping: Throwing AI licenses at the wall and seeing what sticks (or what Velastegui calls the "thousand flowers bloom" approach) isn't leading to tangible returns, she said. * Data: When enterprises are hesitant to give AI agents unfettered access to proprietary data, those agents become less effective, Josh Pantony, CEO of Boosted.ai, which focuses on AI tools for finance, told Axios. What we're watching: Whether companies get more disciplined about AI use. Or overcorrect and clamp down.

Corporations Reeling From Huge AI Costs With No Clear Benefits

Can't-miss innovations from the bleeding edge of science and tech Companies that fell head over heels for AI are experiencing a rude awakening. Costs to access powerful AI tools are soaring, forcing company leaders to ask some difficult questions. As Axios reports, the early warning signs are already here, with Microsoft planning to remove its Anthropic Claude Code licenses after opening up access to the tool just six months ago, reportedly for financial reasons. Uber COO Andrew Macdonald also admitted during a recent podcast appearance that gains in productivity simply weren't being reflected in the company's soaring AI-related expenses. Meanwhile, industry leaders including OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei are walking back their initial claims that AI will lead to a jobs apocalypse, further stoking concerns that the tech may not be all it was initially cracked up to be during the height of the AI hype cycle. It's a perfect storm as companies ponder the real-world benefits from their costly investments in AI, if there even are any. That's particularly true for companies finding that some of their employees are using AI models for meaningless tasks -- like checking the weather, as one CTO told Axios, which is an incredibly expensive and roundabout way of getting a meteorological update. CloudBees CEO Anuj Kapur told the publication that use cases for the tech are limited and that the "reality of AI right now is that it only works for coding." Simply put, many are finding that AI just isn't exactly a money maker. Former Microsoft chief AI officer Sophia Velastegui added that "most people default to automating tasks they dislike rather than tasks most valuable to the company." Then there are ongoing concerns over allowing AI agents to run autonomously could open companies up to new risks, such as data leaks. It's an uncomfortable predicament to be for an AI industry making trillion-dollar bets on imminent surges in demand and soaring revenues. As the Wall Street Journal reported last month, OpenAI missed its own targets of reaching one billion weekly active users for ChatGPT by the end of 2025, as well as several revenue goals. In other words, enterprise customers reeling from soaring costs is the very last thing the AI industry needs. Without meaningful use cases and more clarity on a possible return on investment, firms may think twice before spending vast sums on the tech -- a harsh reality check for an industry that has long heavily relied on hype and seemingly endless investor enthusiasm. More on AI prices: Uber Says Its AI Costs Just Aren't Worth It

Uber burned through its entire 2026 AI budget in four months. Now its COO is questioning whether it's worth it | Fortune

Uber's business model is one of the most AI-forward in Silicon Valley. AI decides your ride price, optimizes your route, among other predictive features. But even with these advanced features, an Uber executive is sounding the alarm on the rideshare company's AI spending. In a recent interview on the Rapid Response podcast, Uber president and chief operating officer Andrew Macdonald said it's hard to draw a connection between the company's rising use of Claude Code and innovations meant to serve consumers. "That link is not there yet," he said. "Maybe implicitly there's more that is getting shipped, but it's very hard to draw a line between one of those stats and 'Okay now we're actually producing like 25% more useful consumer features.'" The comments follow reports that the firm had already burnt through its entire 2026 AI coding tools budget in just four months after incentivizing employees to adopt the technology through an internal leaderboard ranking teams by total AI tool usage. It's the latest development in a complex quandary arising in enterprise AI adoption: increasing AI use comes with higher costs, even as per-unit AI pricing falls. "If you're not actually able to draw a direct line to how [many] useful features and functionality you're shipping to your users, that trade becomes harder to justify," Macdonald said. Uber isn't the only company facing this issue. Microsoft earlier this month reportedly began canceling most of its direct Claude Code licenses, according to The Verge, instead moving engineers toward using GitHub Copilot CLI. A number of other business leaders have walked back their initial bullish AI views. Duolingo CEO Luis von Ahn last year reversed his outlook on AI, saying he doesn't see the tech replacing the tasks his employees perform. Uber didn't immediately respond to Fortune's request for comment. In an earnings call earlier this month, Uber CEO Dara Khosrowshahi said about 10% of the company's committed code is built by autonomous agents. Though he added that the firm's AI use extends beyond its software engineers. "We're seeing uptake of these tools, whether it's our legal team or marketing team or developers," he said. "We think it's creating employees with superpowers." But with more AI use comes higher costs. A recent study from research firm Gartner found that by 2030, inference on highly sophisticated AI models will cost AI firms 90% less than 2025 costs. But the research also found cheaper tokens won't translate to cheaper enterprise AI because agentic models require far more tokens per task than standard models, and because AI providers won't fully pass through lower costs to consumers. Some AI firms are shifting pricing plans to capture increased AI usage. Anthropic changed its pricing model, moving away from a flat fee to a usage-based model, meaning autonomous agents are now charged per token of compute use. In March, OpenAI CEO Sam Altman articulated the industry's broader direction during an interview. "We see a future where intelligence is a utility, like electricity or water, and people buy it from us on a meter," he said. A separate Gartner study forecasts AI agent software spending will reach nearly $207 billion in 2026, up more than 139% from the $86.4 billion spent in 2025. Uber spent 3.4% on research and development in 2025, a 9% increase from 2024. The company spent $951 million on research and development in the first quarter of 2026 alone, a nearly 17% increase from the same time a year ago. Still, Uber isn't shying away from technological innovation. Macdonald said the firm is going all in on autonomous driving, something he sees becoming the norm in under a couple decades. "I don't think my daughters who are little kids today, I don't think they will end up getting a driver's license," he said.

After the AI binge, companies balk at soaring bills

New York (AFP) - Artificial intelligence is getting expensive -- and companies are starting to rethink their embrace of the disruptive technology. Playing by a well-worn Silicon Valley playbook, AI companies charged rock-bottom prices to hook customers after ChatGPT burst onto the scene. Kevin Simback of startup incubator Delphi Labs calls it the era of "subsidized intelligence" -- meaning investors were basically footing the bill so companies could offer AI on the cheap. "But the tides are beginning to turn," Simback warned and an era where the big AI companies actually need to make money has begun -- with leaders OpenAI and Anthropic looking to go public and attract main street investors later this year. Prices are rising across the board, and one big reason is AI agents. Unlike a chatbot that just answers questions, agents actually do things -- book appointments, write code, manage files. And they're expensive to run, because one task can spin up dozens of agents all working at once, each racking up charges. Those charges are measured in tokens -- the basic unit AI companies use to bill customers. A single agent-powered task can burn through dozens of times' more tokens than a simple chat message. Meanwhile, the computer chips and data centers needed to power all this AI can't keep up with demand, creating computing shortages and adding further uncertainty to the nascent industry. "Especially in developer circles, the cost to use AI for things like coding has grown exponentially," said Mark Barton of tech consultancy Omniux. "All the costs are really starting to skyrocket." Some companies have been so eager to use AI that they've gone overboard in a usage binge called "tokenmaxxing." "In some cases people are seeing the cost of tokens exceed the cost of the employee within a month or two of use, just because they're using it too much," says analyst Jack Gold of J.Gold Associates. Smarter spending Even Meta -- which earlier this year encouraged employees to use as many tokens as possible as a measure of productivity -- has had second thoughts. "Nobody should be using AI tools just for the sake of using them," chief technology officer Andrew Bosworth wrote in a memo to staff, reported by the Wall Street Journal. Uber's chief operating officer this week went a step further, raising eyebrows by saying all this AI spending was showing no noticeable increase in productivity. To cut costs, some companies are switching to free, open-source AI models that anyone can download -- not as powerful as ChatGPT or Anthropic's Claude, but good enough for many tasks. Others are moving to smaller, more specialized models built for specific industries like real estate or finance, rather than giant general-purpose ones. And some are simply breaking big AI tasks into smaller steps, handing each piece to the cheapest model that can handle it. The price difference can be dramatic. "The big large monolithic model, it's $15 per million tokens, but you can get that down to like five cents if you use the smaller mini model," says Adrian Balfour of consultancy Enverso. All of this points to AI becoming more like a commodity -- where the specific model matters less than finding the right one at the right price. But don't count out the big players and their state-of-the-art models just yet. "The most advanced users" will always be willing to pay for the best, says John Belton, a portfolio manager at Gabelli Funds.

One company spent half a billion dollars on Claude in a single month: Report comes as AI costs climb

More and more companies are realizing that AI might not be worth its sky-high price tag -- and least, not without limits. One company reportedly learned that the hard way, after its employees blew through $500 million in spending on AI in just one month. An AI consultant told Axios that one of their clients recently spent half a billion dollars in a month on Claude licenses from Anthropic. How did the company rack up such an insanely high bill? Its employees apparently had no limit on how many licenses they could use, leaving them free to splurge on as much Claude as possible. That means expensive AI tools could be deployed for uses humans could easily and quickly accomplish themselves, like checking the weather, as one CTO told Axios their employees were doing. The cautionary tale quickly went viral on social media, with users marveling at how spending half a billion dollars on AI could even be possible. "5 private jets. 2 superyachts. One whole island. Gone. Vaporized into tokens," one user reflected. "Rest in peace to whoever had to send that invoice."

Uber Says Its AI Costs Just Aren't Worth It

Can't-miss innovations from the bleeding edge of science and tech The true cost of AI is rapidly catching up with the tech industry. At first, tech leaders were adamant that their workforce use up as much AI resources as possible, an approach that's become known as "tokenmaxxing." But as prices for cloud AI tools continue to soar, managers are starting to ask pointed questions about whether all of those expenses are actually worth it. Some are even coming to the realization that it may be cheaper to pay human coders after all. In perhaps the most high-profile example of this growing concern yet, Uber COO Andrew Macdonald acknowledged during a recent podcast appearance that gains in productivity simply weren't being reflected in the oodles of cash the company has been shelling out on AI. "That link is not there yet, right?" he told Rapid Response host Bob Safian. "I think maybe implicitly there is more that is getting shipped, but it's very hard to draw a line between one of those stats and, 'Okay, now we're actually producing 25 percent more useful consumer features.'" "If you're not actually able to draw a direct line to how much useful features and functionality you're shipping to your users that trade becomes harder to justify because it's not free," he complained. "AI is not free." While it could "become clearer" over the "coming quarters," Macdonald said that "I think today it's hard even if some of the underlying metrics are trending in a really astronomical direction." During his appearance, Macdonald referenced comments that Uber's CTO Praveen Neppalli Naga made to The Information earlier this year, admitting that the ride hailing app's army of 5,000 engineers had already exhausted the company's 2026 Anthropic Claude Code token budget for the calendar year by mid-March. A reminiscent story is playing out at Microsoft. As The Verge reported earlier this month, the company is planning to remove its Claude Code licenses after opening up access to the tool in December to double down on its in-house Copilot tool instead. While officials maintain the move is meant to streamline operations, employees told the publication that the decision was also financially motivated. Despite these growing concerns, Uber remains all in on AI. Expenses are taking off thanks to investments in AI with CEO Dara Khosrowshahi telling investors during an earnings call earlier this month that the firm was slowing down hiring as a direct result. "We're seeing uptake of these tools, whether it's our legal team or marketing team or developers," he said. "We think it's creating kind of employees with superpowers." In short, Macdonald's comments shows how tech leaders are starting to get antsy about the enormous expenses their companies are shouldering to double down on AI -- and whether they're actually justified. It certainly wouldn't be a shocking revelation if not, given the litany of badly implemented and shockingly unpopular features and glaring bugs caused by faulty AI-generated code.

Uber Boss Seems Unsure About All Those Billions Spent On AI Slop

After years of telling investors and customers that AI was the future, one tech giant leader is sounding a lot less sure about all those burned billions In a recent interview, Uber president and chief operating officer Andrew Macdonald seemed unsure that it made sense for the company to continue to spend billions of dollars on AI tools, such as Claude coding, as there doesn't seem to be a "link" between the rising costs of the tech and actually getting anything useful from it. On May 22, as spotted by The Verge, Macdonald was interviewed by Rapid Response on YouTube. In the interview, the head of the massive ride-sharing and food delivery app explained that so far, there doesn't seem to be much of a direct connection between the ever-rising costs associated with AI tech and the kinds of useful features that can make people more likely to keep using and paying for your app or service. “That link is not there yet, right?" said Macdonald. "I think maybe implicitly there is more that is getting shipped, but it’s very hard to draw a line between one of those stats and, â€~Okay, now we’re actually producing 25 percent more useful consumer features.' I think over the coming quarters and years, maybe that will become clearer, but I think today it’s hard, even if some of the underlying metrics are trending in a really astronomical direction.†In April, Microsoft announced a big change to how it charged people for using GitHub Copilot, its own AI coding tool. Instead of charging a flat monthly fee, it is now going to charge users per token burn. That same month, Anthropic also changed how it charged people to use Claude, its AI coding tool and service. Like Microsoft, people and companies now have to pay based on how often and frequently they use the tool each month. This has quickly made using these AI toolsâ€"and those powered by themâ€"much, much more expensive. Like, using these tools now could cost three times or more than it did before these changes were implemented, back when tech giants were subsidizing the costs and eating billions in losses. “We’re going to have to start talking about token consumption and the associated cost versus headcount,†added Macdonald in the interview. “So if you’re not actually able to draw a direct line to how [many] useful features and functionality you’re shipping to your users, that trade becomes harder to justify.†Keep in mind that it was reported by The Information, and then confirmed by the company, that Uber spent the entirety of its AI budget for 2026 in just about four months. This was because the company didn't anticipate how many people would use it or how expensive it would be to let all of these employees have access to the tech. Uber also spent $3.4 billion on research and development of AI in 2025. In other words, the company is burning billions of dollars on AI, and the guy in charge is going on a podcast and simply shrugging and essentially saying, "I don't know if it's worth it yet." This all seems bad and likely is yet one more sign, among many, that the AI bubble is going to pop in the not-too-distant future. And when it does, a lot of very rich men who are supposedly very smart are going to look really, really dumb. And, folks, I can't wait.

Tokens, tokens everywhere, but not a cent to spare

Companies are facing a significant challenge with the escalating cost of AI tokens as they adopt agentic workflows. Uber has already depleted its annual AI budget, and Salesforce is consuming trillions of tokens. This surge in demand is driving a shift towards usage-based pricing, with projections indicating a massive increase in token usage in the coming years. One of the biggest worries for companies is not just talent or raising the next round, but the rising cost of tokens as they move towards agentic workflows. In the past few weeks, multiple companies such as Uber said that they had spent their entire annual AI budget in a matter of months, while others such as Salesforce are burning trillions of AI tokens. This is occurring even as frontier companies are moving away towards usage-based pricing in line with rising demand. A Goldman Sachs report this May projected agentic AI to raise 24 times to 120 quadrillion tokens per month between 2026 and 2030. ET's Swathi Moorthy looks at the agentic AI landscape and how companies are adapting:

After the AI binge, companies balk at soaring bills

Artificial intelligence is becoming more expensive as companies rethink their initial embrace. The era of "subsidized intelligence" is ending, with rising costs for AI agents and computing power. Businesses are now seeking smarter spending strategies, exploring open-source models and specialized AI to manage expenses. Artificial intelligence is getting expensive -- and companies are starting to rethink their embrace of the disruptive technology. Playing by a well-worn Silicon Valley playbook, AI companies charged rock-bottom prices to hook customers after ChatGPT burst onto the scene. Kevin Simback of startup incubator Delphi Labs calls it the era of "subsidized intelligence" -- meaning investors were basically footing the bill so companies could offer AI on the cheap. "But the tides are beginning to turn," Simback warned and an era where the big AI companies actually need to make money has begun -- with leaders OpenAI and Anthropic looking to go public and attract main street investors later this year. Prices are rising across the board, and one big reason is AI agents. Unlike a chatbot that just answers questions, agents actually do things -- book appointments, write code, manage files. And they're expensive to run, because one task can spin up dozens of agents all working at once, each racking up charges. Those charges are measured in tokens -- the basic unit AI companies use to bill customers. A single agent-powered task can burn through dozens of times' more tokens than a simple chat message. Meanwhile, the computer chips and data centers needed to power all this AI can't keep up with demand, creating computing shortages and adding further uncertainty to the nascent industry. "Especially in developer circles, the cost to use AI for things like coding has grown exponentially," said Mark Barton of tech consultancy Omniux. "All the costs are really starting to skyrocket." Some companies have been so eager to use AI that they've gone overboard in a usage binge called "tokenmaxxing." "In some cases people are seeing the cost of tokens exceed the cost of the employee within a month or two of use, just because they're using it too much," says analyst Jack Gold of J.Gold Associates. Smarter spending Even Meta -- which earlier this year encouraged employees to use as many tokens as possible as a measure of productivity -- has had second thoughts. "Nobody should be using AI tools just for the sake of using them," chief technology officer Andrew Bosworth wrote in a memo to staff, reported by the Wall Street Journal. Uber's chief operating officer this week went a step further, raising eyebrows by saying all this AI spending was showing no noticeable increase in productivity. To cut costs, some companies are switching to free, open-source AI models that anyone can download -- not as powerful as ChatGPT or Anthropic's Claude, but good enough for many tasks. Others are moving to smaller, more specialized models built for specific industries like real estate or finance, rather than giant general-purpose ones. And some are simply breaking big AI tasks into smaller steps, handing each piece to the cheapest model that can handle it. The price difference can be dramatic. "The big large monolithic model, it's $15 per million tokens, but you can get that down to like five cents if you use the smaller mini model," says Adrian Balfour of consultancy Enverso. All of this points to AI becoming more like a commodity -- where the specific model matters less than finding the right one at the right price. But don't count out the big players and their state-of-the-art models just yet. "The most advanced users" will always be willing to pay for the best, says John Belton, a portfolio manager at Gabelli Funds.

Claude AI Bills Getting Out of Control, Company Spends $500 Million in Just One Month

A recent report has brought this issue into sharp focus after an unnamed enterprise reportedly accumulated a staggering $500 million bill in just 30 days through its use of Anthropic's Claude AI platform. The growing adoption of generative artificial intelligence across enterprises is creating a new challenge for business leaders: managing rapidly escalating AI costs. A recent report has brought this issue into sharp focus after an unnamed enterprise reportedly accumulated a staggering $500 million bill in just 30 days through its use of Anthropic's Claude AI platform. The incident is now being widely viewed as a cautionary tale for organizations rushing to deploy AI tools at scale without implementing adequate governance and spending controls. The reported spending surge highlights a growing concern among CIOs, CFOs, and technology leaders worldwide as enterprise AI usage expands from pilot projects to organization-wide deployments. Industry experts note that while generative AI platforms can significantly improve productivity, costs can rise dramatically when thousands of employees simultaneously access premium AI models without governance mechanisms in place. Enterprise AI Faces Its First Major Cost Reckoning Microsoft and Amazon Tighten AI Controls The growing focus on AI costs is not limited to a single organization. The Rise of AI Governance A Warning for the Enterprise AI Era