The Trillion-Dollar Bet: Inside Big Tech’s Debt-Fueled AI Buildout

Hyperscalers have issued $121 billion in bonds this year alone to finance data centers and GPUs—but with ROI timelines stretching beyond 2030, the largest infrastructure bet in corporate history is raising questions Wall Street can’t ignore.

Hyperscalers—massive cloud Infrastructure providers such as Amazon, Alphabet/Google, Meta, Microsoft, and Oracle—have issued roughly $121 billion in new debt so far in 2025, with over $90 billion raised in the past three months alone, according to analysis from Mellon Investments. Hyperscalers have added $121 billion in new debt this year—more than four times the average annual issuance over the previous five years, per Bank of America data.

The spending surge is transforming historically cash-rich businesses into leveraged operators. Hyperscalers are increasingly leaning on debt markets to bridge the gap between rapidly rising AI Capex budgets and internal free cash flow, transforming historically cash-funded business models into ones utilizing leverage. Aggregate capex for ‘the big five’, after buybacks and dividends are included, are now above projected cash flows, thereby necessitating external funding needs, according to MUFG Americas.

Capital Expenditure Approaching $700 Billion

The numbers behind the AI arms race defy historical comparison. Hyperscalers, including Amazon, Microsoft, Meta and Alphabet, announced capital expenditure could hit $700 billion on AI this year, according to CNBC analysis—more than the GDP of Singapore or Israel.

Amazon said on Thursday it would invest about $200 billion in capital expenditures in 2026, an announcement that followed Alphabet telling investors on Wednesday its capex would fall between $175 billion and $185 billion this year. Late last month, Meta told investors it would spend anywhere from $115 billion to $135 billion in 2026, while Microsoft’s annual run rate for its 2026 fiscal year would put the company on pace for capital expenditures of $145 billion, per Yahoo Finance.

Hyperscaler capex for the ‘big five’ (Amazon, Alphabet/Google, Microsoft, Meta/Facebook, Oracle) is now widely forecast to exceed $600 bn in 2026, a 36% increase over 2025. Roughly 75%, or $450 bn, of that spend is directly tied to AI infrastructure (i.e., servers, GPUs, datacenters, equipment), rather than traditional cloud, according to IEEE ComSoc Technology Blog analysis.

Free Cash Flow Under Pressure

The spending surge is decimating free cash flow across the sector. Last year, the four biggest U.S. internet companies generated a combined $200 billion in free cash flow, down from $237 billion in 2024. The more dramatic drop appears to be ahead, as companies invest heavily up front, promising future returns on investment, according to CNBC.

Analysts at Barclays now see a drop of almost 90% in Meta’s free cash flow, after the social media company said last week that capex this year will reach as high as $135 billion. They kept their overweight rating even as they forecast an even tougher cash position the next two years. ‘We are now modeling negative FCF for ’27 and ’28, which is somewhat shocking to us but likely what we eventually see for all companies in the AI infrastructure arms race,’ the analysts wrote, per CNBC.

Amazon, which on Thursday said it expects to spend $200 billion this year, is now looking at negative free cash flow of almost $17 billion in 2026, according to analysts at Morgan Stanley, while Bank of America analysts see a deficit of $28 billion. In a filing with the SEC on Friday, Amazon let investors know that it may seek to raise equity and debt as its build-out continues, according to CNBC.

Pivotal Research projects Alphabet’s free cash flow to plummet almost 90% this year to $8.2 billion from $73.3 billion in 2025, according to CNBC reporting.

The ROI Question: When Will Returns Materialize?

The critical weakness in the AI infrastructure thesis is the uncertain timeline to profitability. Nearly two-thirds of companies (63%) are actively measuring return on investment—and they’re projecting it will take an average of 28 months to recover their upfront costs and realize meaningful returns, according to Gallagher’s 2026 AI Adoption and Risk Survey.

The Think Circle report highlights that although many executives are investing in AI, few can reliably measure ROI today—with only about 29% saying they can measure ROI confidently. An IBM CEO study found that only around 25% of AI initiatives deliver expected ROI, just 16% have scaled enterprise-wide, and CEOs are balancing pressure for short-term ROI with longer-term innovation goals, per IBM research.

Timelines on recouping huge capex are ‘very much unknown’ right now. ‘The estimated useful life on much of this spend, including data centers and chips, can be as low as 3-5 years, meaning the hyperscalers need to see significant returns on investment before 2030—providing a very tight timeline,’ according to CNBC analysis citing Third Bridge analyst Charlie Field.

Bain & Co has calculated that with the growth rate of AI compute demand running at more than double Moore’s Law, the industry is not going to be able to fund the required capex. This level of growth would require annual investment of around $500 billion. To fund that, AI companies would have to generate $2 trillion in revenue, but Bain estimates that not even the most aggressive take-up of enterprise AI would raise $1.2 trillion, leaving an $800 billion spending gap. This methodology does not include revenue from consumer AI services—but $800 billion is a lot of subscriptions—or ads, per IEEE ComSoc reporting.

Echoes of the Dot-Com Era

The scale and financing structure of AI infrastructure spending has drawn comparisons to the telecom bubble of the late 1990s. In the five years after the American Telecommunications Act of 1996 went into effect, telecommunications equipment companies invested more than $500 billion, mostly financed with debt, into laying fiber optic cable, adding new switches, according to Wikipedia.

A more recent and relevant cautionary tale is the telecom bubble of the late 1990s. In anticipation of explosive internet growth, companies like Global Crossing and AT&T spent over $500 billion laying an estimated 80 million miles of fiber optic cable. With as much as 85% of the new fiber left unused, the resulting capacity glut caused the cost of bandwidth to plummet by 90%, per WinBuzzer analysis.

Parallels to the dot-com and telecom buildouts suggest risk if infrastructure returns lag expectations, according to Mellon Investments.

Oracle: The Canary in the Coal Mine

Oracle’s aggressive AI buildout illustrates the financial stress points. Barclays predicts Oracle may run out of cash by November 2026 if the current trajectory continues. Starting FY 2027 (June 2026), Oracle faces financing gaps that will re