Big Tech's AI ambitions fuel $300 billion borrowing boom as Wall Street shows investor fatigue

How Big Tech's AI Ambitions Are Fueling a Borrowing Boom

For the past few years, the largest US technology companies have been in a costly race to develop advanced artificial intelligence systems while at the same time providing computing power to a burgeoning field of startups. To chase these goals, they have radically changed how they finance their growth. Long reliant on rich revenues and share price increases, Alphabet Inc.'s Google, Meta Platforms Inc. and other tech giants now are borrowing heavily to build the technology that makes chatbots run. In March, Amazon.com Inc. sold bonds in Europe for the first time, raising €14.5 billion ($17 billion), the biggest ever corporate deal in the currency. The retail giant also borrowed $37 billion in the US bond market, in the fourth-largest US corporate bond sale on record. And Facebook parent Meta sold $25 billion of investment-grade bonds on April 30 aimed at financing its AI buildout, Bloomberg reported. Four of the biggest US tech companies have said that just this year, they need to spend around $650 billion collectively on data centers, networking equipment and other AI infrastructure needed to achieve their AI ambitions. Here's how the reliance on borrowing is changing the tech industry and fueling the AI boom. How has the growth of AI changed tech companies' financial practices? For many years, tech companies that emerged from the internet boom grew by reinvesting their immense profits back into their businesses. They also sold bonds, but that played a smaller role in raising and spending capital. But beginning in late 2025, big tech firms began issuing tens of billions of dollars in debt as they scrambled to ramp up investing in AI capacity. New businesses such as OpenAI and Anthropic, meanwhile, have each raised billions of dollars in funding from venture capital investors. How are tech companies planning to spend the money? Most of the money these tech companies are deploying -- be it through debt or equity -- is being invested in AI-related equipment, services and real estate. Alphabet alone said that about 40% of its technical infrastructure spending was tied to data centers and networking gear, while 60% was tied to servers. Oracle Corp. is a prime example of the data center expense. The database giant has been raising both corporate debt and loans for specific projects to finance the development of data centers across the country. It's not just real estate, however. These firms also need to stock the facilities with expensive chips to train and run AI models. Often, companies create special purpose vehicles -- essentially, separate businesses launched for specific financial goals including acquiring tech gear. With an SPV, the debt is kept off the company's balance sheet, protecting the business from potential downgrades. Since late 2025, Elon Musk's xAI has been working on raising as much as $20 billion via off-balance-sheet vehicles that buy chips and lease them back to xAI. Two other expenses compound the race to be first: power costs and AI talent. Alphabet recently purchased a clean energy developer to power its data centers as the US electricity grid struggles to meet the projects' demands. Meta also has been spending millions of dollars to hire skilled engineers. Why are companies borrowing instead of using cash or issuing shares? The pressure on big tech companies to build out data centers to power AI functions is immense. Meta, Alphabet and the other tech heavyweights can use the existing cash in their coffers for their data center buildouts. Their advertising businesses give them enough cash that they can borrow comfortably and also reinvest some of that revenue into AI. Google, for instance, took in more than $97 billion in revenue, excluding partner payouts, in the fourth quarter of 2025. But borrowing can still be attractive, especially at a time when Wall Street firms are eager to lend them money. The special-purpose entities that enable companies to keep debt off their balance sheets add to the appeal. Borrowing heavily isn't always an option for AI startups, which typically have far less revenue. Instead, privately held firms such as OpenAI and xAI have raised billions of dollars by selling stakes in their companies and using that cash for their AI needs. But they can only do that so many times, as equity holders see their stakes get increasingly diluted. In 2025 xAI borrowed $5 billion in corporate debt, which the startup has since repaid. OpenAI and Anthropic haven't tapped the debt capital markets yet and are looking at other ways of raising money. How unusual is this level of borrowing -- and what's different about this moment? The wave of AI-related borrowing spooked investors late last year, when large technology companies raised nearly $100 billion within a few weeks to expand cloud and data-center capacity. The surge came on the heels of a roughly $30 billion financing tied to the construction of a Meta data center in Louisiana, a transaction that underscored both the scale of capital required for AI infrastructure and the increasingly varied ways companies are structuring the funding. That deal was raised by a special purpose vehicle from Meta but will pay back lenders through a long-term lease with the tech giant. The structure showed how data center operators can issue traditional bonds while also raising large sums from lenders without adding too much debt to their balance sheets and jeopardizing their credit ratings. Underscoring the thirst for capital, Alphabet sold a rare 100-year bond in early 2026, a type of deal that no tech company had done since the late 1990s and that tapped demand from long-term investors such as insurance and pension funds. And Meta returned to the debt markets for its bond sale of $25 billion in investment-grade bonds. The debt sale came a day after Meta projected higher capital expenditures for the year than it had forecast in January. The AI borrowing wave also stands out for its speed, size and for the type of borrowers involved. Surges in corporate debt have historically been associated with speculative bubbles, such as during the 1980s leveraged buyout boom, when high-risk bonds were issued to finance a wave of aggressive corporate acquisitions. By contrast, the recent issuance has come from some of the most cash-rich and highly rated companies in the world. How does taking on this much debt change the risk profile of these companies? Tens of billions of dollars have been raised in compressed time frames, and the AI buildout is proceeding despite elevated interest rates, reflecting the competitive urgency of generative AI. Some market participants have drawn parallels to earlier infrastructure booms, such as the fiber-optic buildout during the dot-com era, when telecom companies borrowed heavily to lay network capacity. But there are key differences. Today's largest issuers are far more profitable and diversified than many telecom operators were in the late 1990s. Despite the recent surge, debt financing is projected to remain a relatively small share of total AI spending by large technology companies. Analysts have estimated that approximately 80%-90% of their planned capital expenditures are expected to be funded through operating cash flows. And despite the recent uptick in borrowing, major data center operators are anticipated to maintain low levels of total debt compared with their annual earnings. Get the Tech Newsletter bundle. Get the Tech Newsletter bundle. Get the Tech Newsletter bundle. Bloomberg's subscriber-only tech newsletters, and full access to all the articles they feature. Bloomberg's subscriber-only tech newsletters, and full access to all the articles they feature. Bloomberg's subscriber-only tech newsletters, and full access to all the articles they feature. Plus Signed UpPlus Sign UpPlus Sign Up By continuing, I agree to the Privacy Policy and Terms of Service. Still, the sheer magnitude of the financing has implications. Higher borrowing could alter companies' financial profiles, influencing their credit ratings and their ability to borrow cheaply. Beyond the tech sector, heavy debt issuance also might reshape credit markets by absorbing investor demand that otherwise might have flowed to other industries. This could raise borrowing costs for other companies, too, while increasing lenders' exposure to an industry whose long-term returns on AI investments are still being proven. Morgan Stanley in 2025 forecast that issuance of investment-grade corporate bonds in 2026 could top $2 trillion, a record, driven in part by artificial intelligence investments. Markets for high-grade bonds could need to absorb about $1.5 trillion in AI data center bond issuance over the next five years, JPMorgan analysts estimated last year. By 2030, that debt could account for more than 20% of the investment-grade bond market, they said. What could go wrong if the AI boom doesn't meet expectations? If the AI boom doesn't live up to expectations, big tech companies that have been spending heavily on data centers, chips and power to support AI could end up with too much capacity and equipment that becomes outdated quickly. That would be similar to what happened during the dot-com bubble, when telecom firms built more network capacity than customers needed. Lower-than-expected profits would squeeze companies' cash flows, potentially forcing businesses to cut investments or borrow more, eventually weakening their financial profiles. There are also broader market risks. Investors have poured money into tech debt and shares, betting on AI-driven growth. If that optimism fades, stock prices could fall and lenders would take a big hit.

Investors push for higher yield on $14bn of Oracle-backed data centre debt

Investors are demanding a higher-than-usual premium to back a $14bn bond offering by an Oracle-backed data centre project, amid growing concerns over the huge amounts of AI-related debt coming on to the market. The sale, which has been privately pitched to a small group of institutions in recent weeks, is to fund a 1 gigawatt data centre in Saline Township, Michigan, as part of a $300bn agreement with OpenAI to provide the ChatGPT maker with 4.5GW of computing power, say people familiar with the matter. However, some investors have questioned whether Oracle -- whose aggressive AI spending plans have been a source of market worry in recent months -- could provide sufficient guarantees to ensure debt repayment if the project were delayed or if Oracle ever exited the lease backing the project bond. The tech giant is only a tenant, rather than a joint owner of the facility. To compensate for these risks, some investors had demanded yields of more than one percentage point over Oracle's publicly traded corporate bonds due in 2040, the people said, and more credit protections, including the removal of a proposed option that would allow the issuer to buy back the debt early. Few, however, expect this ultimately to scupper the deal as the order book is still well covered. "It's still project finance ultimately," one of the people said. "You're still going to take construction risks." Big Tech companies have borrowed more than $100bn in global bond markets so far this year to help fund the AI arms race, fuelling investor concerns that the runaway capital expenditure might not translate into actual profits and may be a bubble that could hit the broader economy. The Oracle-backed data centre is being built by developer Related Digital, and the debt is being issued by a special purpose vehicle. Investors say that raising debt off Oracle's balance sheet helps alleviate their concerns over the rising debt load at the company, which raised $25bn from the bond market in February after it pledged to preserve its investment-grade credit rating. However, if the deal went ahead, it it could open up a new wave of financing for similar project-level bonds without equity stakes from big tech companies, said Gianluca Bacchiocchi, a partner at Clifford Chance specialising in financing energy and infrastructure projects. The sale was being led by Bank of America and was expected to launch in the so-called 144A market in the coming weeks, allowing thousands of institutional investors to trade the bond, the people said, adding that final pricing and terms were still subject to negotiation. Some investors who had been approached privately were hoping to secure the debt early on and resell a portion of it for a quick profit when the deal launched to the wider group of investors, the people said. "This is a new asset class that only emerged in recent months. There is no quant model for us to calculate what is a good relative value for a deal," said Katie DeSplinter, head of US credit trading at Capital Group, speaking about such deals in general. Pimco was expected to be an anchor investor of the bond sale, while Blackstone would provide about $2bn of equity in addition to the debt, the people added. Blue Owl Capital also considered backing the project but did not proceed as negotiations stalled, the FT previously reported. Bank of America, Related Digital, Blackstone and Pimco declined to comment. Oracle did not respond to requests for comment. Unlike traditional project loans that need to be refinanced within four to six years, a project bond can provide financing with longer maturities to minimise refinancing risks. While banks used to provide the bulk of financing for data centre build-outs, large-scale projects are increasingly turning to other types of debt investors as the massive deal sizes overwhelmed the bank market. "There's more caution among banks," said Bacchiocchi. "Some are reviewing whether they need to reduce or hedge some of their positions." Additional reporting by Eric Platt and Kate Duguid in New York

AI Debt Investors Show Signs of Fatigue After $300 Billion Binge

After a $300 billion AI debt binge that spanned every corner of the credit market, investors are starting to show some signs of fatigue. Make no mistake: There's still appetite for those deals. But bankers have had to work harder to sell them lately, offering more incentives and higher compensation to investors who are spoiled for choice. Just on Thursday, they had the option of buying into Meta Platforms Inc.'s jumbo investment-grade offering, or take more risk and look at the latest leveraged loan deal from CoreWeave Inc. But there are signs of waning enthusiasm. Meta racked up a peak order book of about $96 billion for a bond sale that's expected to raise as much as $25 billion, according to people with knowledge of the matter. The last time it tapped the corporate bond market in October, it drew $125 billion of demand for a $30 billion deal. Get the Tech Newsletter bundle. Get the Tech Newsletter bundle. Get the Tech Newsletter bundle. Bloomberg's subscriber-only tech newsletters, and full access to all the articles they feature. Bloomberg's subscriber-only tech newsletters, and full access to all the articles they feature. Bloomberg's subscriber-only tech newsletters, and full access to all the articles they feature. Plus Signed UpPlus Sign UpPlus Sign Up By continuing, I agree to the Privacy Policy and Terms of Service. And lately, investors are demanding more protection. Borrowers are agreeing to repay some or all of the principal on debt before it matures, a clause called amortization that removes risk down the line. More issuers -- especially in the junk-bond market -- are getting so-called backstops from Alphabet Inc.'s Google, a covenant guaranteeing that a data center lease will be paid even if a tenant defaults. There are also provisions that place a ceiling on building costs. It all suggests an undercurrent of anxiety around a market that has held up well so far -- but is still relatively new and untested. And with estimates that the AI buildout could cost some $3 trillion, the size of the borrowings is only likely to balloon from here. "At the end of the day, these companies are selling a lot of debt and they're going to have to pay up to borrow," said Robert Tipp, head of global bonds at PGIM Fixed Income. "The market, after a spectacular narrowing in corporate spreads to historical tights, is seeing a wall of worry piled up before it." Many money managers say there aren't always conventions for how provisions designed to give lenders more protection should translate to yields and risk premiums on the securities. "We're seeing what different investors value when it comes to these financings and how they're evaluating risk and return," John Servidea, global co-head of investment grade debt capital markets at JPMorgan Chase & Co. said about data-center debt. "We're seeing really good demand for these deals but as supply increases, we expect deal terms and structures to continue to evolve." And over time, debt buyers have started to get a better grasp of the various risks involved in data-center projects, according to David Kinsley, a senior portfolio manager at Impax Asset Management. He says investors are asking questions like: "Do you have execution issues? Supply chain or construction delays? Is it a lower-quality tenant?" They're also turning away from deals that don't meet their standards. One investor, speaking anonymously to discuss private deliberations, said they passed on a $14 billion offering for an Oracle Corp. data center in Michigan partly because the bonds are callable. That adds risk for creditors since a borrower can refinance at a lower interest rate once construction is complete. Big Tech Effect The need for due diligence is even more important in the riskier parts of the credit market. One thing that helps to put investors at ease is having a connection to a hyperscaler -- a company with a massive cloud computing operation -- either as a tenant or a backstopper. Of the roughly $27 billion in high-yield debt issued this year to fund AI infrastructure, almost a third of that money is tied to the most valuable publicly-traded company in the world, Nvidia Corp. An entity backed by asset manager Tract Capital has raised more than $8 billion since Febraury to finance construction of a data center in Nevada, where the chipmaker will be a tenant. Partnering with a tech giant can also help data center developers obtain investment-grade ratings. Earlier this month, Blackstone Inc.-backed QTS sold $4.6 billion of bonds to fund a Microsoft Corp.-tied facility. Moody's Ratings gave a first-time credit score of Baa2 to the offering -- two notches above junk. The tech behemoths aren't showing signs of slowing down. Four hyperscalers reported earnings on Wednesday, and they all reiterated their intentions to spend big on data centers and the chips that power them. Still, there are multiple factors for investors to consider when looking to buy data-center debt, and having a hyperscaler involved doesn't mitigate all of those risks. "All data center credits are not created equal," said Grant Nachman, Chief Investment Officer at Shorecliff Asset Management. "Future tenant quality alone isn't sufficient. Bondholders need confidence the borrower can build and power the asset, and keep it online. The safest deals meet multiple criteria, and are structured to" pay down debt quickly.

Oracle's $16.3B data centre financing required PIMCO to anchor $10B after US banks retreated

In short: Oracle closed a $16.3 billion financing for a single data centre campus in Saline Township, Michigan, the largest single-facility technology debt package ever assembled. PIMCO anchored roughly $10 billion of the bond tranche because US banks retreated from the deal, citing doubts about AI infrastructure demand sustainability. The financing is part of $72 billion in total data centre partner debt Oracle has assembled across Michigan, Texas, Wisconsin, and New Mexico for the Stargate joint venture, all backed by a company with BBB-negative credit outlook and $553 billion in performance obligations concentrated heavily in a single counterparty, OpenAI. Oracle's $16.3 billion data centre financing closed this week, the largest single-facility technology debt package ever assembled, and it required a bond fund to anchor it because America's banks were not willing to. PIMCO, the world's largest active fixed-income manager, purchased approximately $10 billion of the roughly $14 billion bond tranche, according to Bloomberg. The remainder of the capital stack includes approximately $2 billion in equity from Related Digital Infrastructure and Blackstone. Bank of America structured the deal. Goldman Sachs and Wells Fargo advised Related Digital. The financing covers a single data centre campus in Saline Township, Michigan, designed to exceed one gigawatt of capacity and serve as infrastructure for Oracle and OpenAI's Stargate project, the joint venture announced in January 2025 with SoftBank that has become the gravitational centre of American AI infrastructure spending. The bonds carry a 7.5% coupon with a 19.5-year maturity, structured with six years of interest-only payments followed by 13 years of amortisation. This is not cheap debt. Oracle's own corporate bonds trade at significantly tighter spreads. The premium reflects the project finance nature of the deal: the bonds are secured against the Michigan campus itself, not against Oracle's corporate balance sheet, which means investors are lending against a single facility's projected cash flows rather than the creditworthiness of a $400 billion enterprise. The structure isolates the risk but prices it accordingly. At 7.5% over nearly two decades, the total interest cost will exceed the principal, a financing burden that only makes economic sense if the facility generates revenue from day one and maintains occupancy for the duration of the loan. Oracle's $553 billion in remaining performance obligations, reported in Q3 of fiscal 2026, provides the demand signal that underwrites that assumption, but the concentration risk is considerable. A significant share of those obligations traces back to a single counterparty: OpenAI. TD Cowen reported earlier this month that US banks have been retreating from Oracle data centre financing, citing doubts about the sustainability of AI infrastructure demand at the scale Oracle is projecting. That retreat is why PIMCO ended up anchoring $10 billion of the bond offering. When traditional lenders step back from a deal, it does not necessarily mean the deal is bad. It means the risk-reward calculation has shifted to a point where banks, constrained by regulatory capital requirements and concentrated exposure limits, cannot justify the position. Asset managers like PIMCO, which manage discretionary capital with longer time horizons, can. But the signal matters. The largest data centre financing in history required the bond market to do what the banking system would not. The Michigan campus is not Oracle's only data centre megaproject. It is not even the largest. Oracle has assembled at least $72 billion in total data centre partner debt across three major financing packages: the $16.3 billion Michigan deal, approximately $38 billion for campuses in Texas and Wisconsin, and roughly $18 billion for a facility in New Mexico. All of these are being built to serve the Stargate joint venture and Oracle's broader cloud infrastructure expansion. Oracle's own capital expenditure for fiscal 2026 is expected to reach approximately $50 billion, more than double the prior year. Oracle's $50 billion capex programme, outlined by CFO Hilary Maxson, represents the company's largest-ever infrastructure investment cycle and has strained the balance sheet to a degree that prompted both S&P and Moody's to assign negative outlooks to Oracle's BBB and Baa2 credit ratings respectively. UBS analysts have said a downgrade to junk is unlikely, but the trajectory is clear: Oracle is leveraging its investment-grade rating to the maximum extent the market will tolerate. The sheer volume of capital flowing into AI data centres has created a parallel financial infrastructure. Jane Street's involvement in CoreWeave's $6 billion debt facility and $1 billion equity raise demonstrated that Wall Street's most sophisticated trading firms see AI cloud infrastructure as a fixed-income opportunity. Blackstone's $10 billion data centre debt financing in Australia through Firmus showed the same pattern repeating on another continent. Bain Capital's $5 billion stake sale in Bridge Data Centres confirmed that private equity sees data centre infrastructure as a liquid, tradeable asset class. Oracle's Michigan deal is the largest single transaction in this emerging market, but it is part of a pattern: AI infrastructure is being financed like real estate, with project-level debt secured against physical assets and long-term lease obligations, because the capital requirements have exceeded what corporate balance sheets alone can support. Oracle's transformation from a database and enterprise software company into an AI infrastructure provider has been the most dramatic strategic pivot in enterprise technology since Microsoft's cloud transition under Satya Nadella. Oracle's cloud revenue has been growing at more than 20% annually. Its remaining performance obligations of $553 billion in Q3 fiscal 2026 represent contracted future revenue that provides visibility into demand. But contracted revenue is not the same as collected revenue, and the gap between signing a cloud contract and building the infrastructure to fulfil it is where Oracle's financial risk concentrates. The company announced 30,000 layoffs earlier this year, a workforce reduction that executives framed as a reallocation of resources from legacy operations to cloud and AI infrastructure. The human cost of the pivot is being borne by employees whose skills were built for a business Oracle is leaving behind. The $1.4 trillion in utility capital expenditure projected through 2030 to support AI data centre power demand illustrates the infrastructure chain that must function for Oracle's bet to pay off. A one-gigawatt data centre campus requires dedicated power generation, transmission upgrades, water cooling systems, and grid interconnection agreements that take years to complete. Saline Township is in Michigan, where DTE Energy provides electricity and where the state has been courting data centre development with tax incentives and expedited permitting. But power availability is the binding constraint on data centre expansion nationally, and every gigawatt committed to Oracle is a gigawatt unavailable to competitors. The fundamental question the Oracle financing raises is not whether AI infrastructure demand is real. It is whether the demand is durable enough to justify $72 billion in project-level debt with 19.5-year maturities. The history of technology infrastructure is littered with overbuild cycles: the fibre-optic boom of the late 1990s, the data centre construction surge of 2006 to 2008, the first wave of cloud buildouts that left Amazon Web Services with years of excess capacity before demand caught up. In each case, the technology ultimately justified the investment, but the companies and creditors who financed the first wave often did not survive to benefit. The fibre laid in 1999 carries today's internet traffic, but the companies that laid it went through bankruptcy first. Oracle's counterargument is that this cycle is different because the demand is contractually committed, not speculative. The Stargate partnership with OpenAI and SoftBank provides a named customer with binding obligations. But OpenAI itself is a company that has never generated an annual profit, is burning cash at an accelerating rate, and is in the process of converting from a nonprofit to a for-profit structure in a transaction that remains legally contested. The durability of Oracle's $553 billion in performance obligations depends on the financial health of the customers who signed them, and the largest of those customers is itself dependent on continued venture capital and strategic investment to fund its operations. PIMCO's willingness to anchor $10 billion of the Michigan financing suggests that the world's most sophisticated bond investors believe the risk is manageable. The retreat of US banks from the same deal suggests that not everyone agrees. The $16.3 billion got over the line. The question is whether, 19.5 years from now, the campus it financed will look like a prescient bet on the defining technology of the century or like another monument to a cycle that peaked before the debt was repaid.

Why Oracle's AI Spending Spree Has Wall Street On Edge - Oracle (NYSE:ORCL)

Debt Limits Challenge Data-Center Financing The report said Oracle's $300 billion megadeal with OpenAI is pushing Wall Street to its limits, straining investor appetite for the massive debt fueling America's data-center boom. Analysts Warn Of Large Funding Needs Morgan Stanley credit analysts told the WSJ that Oracle still faces more than $100 billion in additional funding needs through 2027 and early 2028, after planning to raise about $50 billion for 2026. They cautioned that these requirements could "test the depths of different fixed-income markets," underscoring the scale of financing needed for AI infrastructure. Market Risks And Investor Concerns Grow Lenders have grown cautious due to Oracle's weaker financial profile compared with peers, including higher debt and cash burn, as well as its reliance on OpenAI, which adds uncertainty. Despite Oracle stating that projects are progressing on schedule and funding sources are diversified, the broader market faces pressure as AI spending outpaces available capital, with big tech expected to fund only about half of the projected $3 trillion investment through 2028, according to Morgan Stanley. Execution Remains The Key Risk He noted Oracle often serves as an overflow provider but remains much closer to these leaders than smaller cloud players, supported by its software cash flows and an estimated $550 billion backlog that signals strong demand visibility. However, Luria warned that execution will be critical as Oracle scales its infrastructure. While earlier concerns about AI demand have eased -- supported by large funding commitments like OpenAI's $122 billion raise and ongoing compute shortages -- he said the company's ability to deliver on this opportunity remains the key risk. ORCL Price Action: Oracle shares were up 1.52% at $178.96 during premarket trading on Friday, according to Benzinga Pro data. Photo via Shutterstock Market News and Data brought to you by Benzinga APIs To add Benzinga News as your preferred source on Google, click here.