Aggressive artificial intelligence spending by Microsoft, Amazon, Alphabet, and Meta is rapidly depleting free cash flow, with analysts projecting aggregate losses to reach $64 billion in 2026. Despite ballooning capital expenditures and negative cash flows at major tech firms, executives defend the investments as unprecedented demand for cloud infrastructure continues to accelerate.
The massive infrastructure race powering artificial intelligence has crossed a critical financial threshold, directly impacting free cash flow metrics across Wall Street. Following second-quarter earnings reports from Microsoft, Amazon, Alphabet, and Meta, analysts at Bank of America raised their collective capital spending outlook for hyperscalers to roughly $860 billion for 2026, with projections climbing toward $1.2 trillion in 2027.
That staggering financial acceleration is altering corporate balance sheets. Alphabet reported that its free cash flow turned negative during the second quarter for the first time since its public offering in 2004, while Amazon’s trailing 12-month free cash flow slipped into negative territory amid surging capital expenditures. Meta maintained a barely positive quarterly cash flow, and Microsoft remained firmly in the black.
Amazon Sparking Wall Street Alarm With a $200 Billion Capex Plan
The scale of the spending came into sharp focus when Amazon announced plans for $200 billion in capital expenditures for the year, blowing past Wall Street expectations by more than $50 billion. That outsized guidance triggered a more than 10% drop in Amazon shares during extended trading, compounded by a profit forecast that also missed analyst consensus.
During an earnings call with executives, Evercore ISI tech analyst Mark Mahaney pressed leadership on the timeline for returns. The strong long-term return on investing capital — I think that’s the debate in the market today,
Mahaney said during the call, asking how investors would be able to track those returns.
Amazon CEO Andy Jassy defended the strategy, arguing that new artificial intelligence capacity is being monetized as quickly as it goes live. Jassy categorized the buildout as a very unusual opportunity
that is accelerating customer migration to cloud services while helping AWS forecast demand efficiently.
Bank of America Forecasts Aggregate Free Cash Flow Deficits Through 2028
For most of the past decade, the major cloud providers—including major US hyperscalers alongside Alibaba, Tencent, and Baidu—historically generated between $135 billion and $272 billion in annual free cash flow, maintaining margins between 10% and 20%. Bank of America projections show that traditional cash generation model turning upside down.
Analysts forecast that the combined free cash flow across the eight tracked companies will plunge from an estimated $180 billion in 2025 to roughly -$64 billion in 2026. The deficit is expected to widen further to -$144 billion in 2027 and -$186 billion in 2028, with aggregate free cash flow margins dropping to -2.8% this year and sliding to -5.8% by 2028.
| Metric / Year | 2025 (Est.) | 2026 (Proj.) | 2027 (Proj.) | 2028 (Proj.) |
|---|---|---|---|---|
| Aggregate Free Cash Flow | $180 Billion | -$64 Billion | -$144 Billion | -$186 Billion |
| Aggregate FCF Margins | Positive | -2.8% | -5.4% | -5.8% |
| Hyperscaler Capex (Four US Firms) | — | ~$800B – $860B | Up to $1.2 Trillion | — |
Despite the negative cash flow trajectory, financial analysts see substantial offsetting momentum. Google posted an 82% year-on-year revenue increase in its cloud division during the same quarter it posted negative free cash flow. Across the largest cloud providers, customer commitments and remaining performance obligations now total approximately $2.3 trillion.
Jassy and Bank of America Analysts Weigh Execution Risks Versus Surging Demand
Financial institutions continue to question whether companies maintain adequate safeguards against overbuilding. JPMorgan analyst Doug Anmuth specifically asked Amazon leadership whether the corporation had financial guardrails
in place to manage the unprecedented capital outlay.
Jassy maintained that the aggressive posture is necessary to capture market share, noting that Amazon’s proprietary in-house chips, Trainium and Graviton, are on track to generate more than $10 billion in revenue. He also characterized the artificial intelligence market as barbelled,
anchored by heavy spending from major AI labs on one end and enterprise productivity tools on the other, leaving a massive middle tier of production workloads still waiting to be deployed.
Bank of America analysts echoed the tension between aggressive deployment and financial health, pointing out that execution remains the primary variable. A key risk remains execution … but we see surging demand outpacing the capacity being built resulting in sustainably strong demand in the mid-term,
the analysts wrote, adding that aggregate annual free cash flow should eventually exceed historical averages once the infrastructure is fully operational.
Worth a look