Markets Brace for $65 Billion in AI Capital as Google and Anthropic Prepare Raises
Two buses arrive at once, and the equity market's appetite appears to hold
Photo by Christian Wiediger on Unsplash
Google's $40B Anthropic investment and Amazon's $25B pledge test whether markets can absorb unprecedented private AI funding without crowding out other…
The Two-Bus Problem
There's an old London saying about buses: wait forever, then two arrive at once. That's the situation equity markets face as Google and Amazon lock in capital commitments to Anthropic that, combined, approach $65 billion in pledged funding.
Google's structure follows a now-familiar playbook. The initial $10 billion arrives at a $350 billion valuation, with an additional $30 billion tied to performance milestones. Amazon, days earlier, committed $5 billion upfront with provisions for up to $20 billion more. The capital doesn't flow all at once. But the pledges signal a willingness by two of the largest balance sheets in tech to deploy vast sums into a single private company, and markets have responded by bidding Alphabet shares to session highs.
What's notable is how little resistance the raises have met. Credit spreads remain tight. The S&P hasn't buckled under the weight of capital reallocation. Either markets believe the money will cycle back quickly through cloud revenue, or they've concluded that AI infrastructure investment is now a permanent fixture of corporate capital allocation. Probably both.
Capital Recycling and the Hyperscaler Model
The deals are structured so that much of the capital flows back to Google and Amazon in the form of cloud compute purchases. Anthropic has secured 5 gigawatts of TPU capacity from Google and another 5 gigawatts from Amazon's AWS. For context, 5 gigawatts is roughly the peak summer electricity load of metropolitan San Francisco. It's a staggering infrastructure commitment.
The circular nature of these agreements is the tell. Alphabet invests cash into Anthropic; Anthropic spends that cash (and more) on Google Cloud TPU capacity; Alphabet books the cloud revenue and marks up its equity stake. Amazon runs the same loop through AWS. When both hyperscalers reported Q1 earnings, significant portions of their net income gains traced not to operational improvement but to marking up Anthropic equity. Alphabet's $62.6 billion quarterly profit included roughly $28.7 billion from private company revaluations, primarily Anthropic. Amazon disclosed $16.8 billion in gains from the same source.
This accounting quirk muddies the picture for anyone trying to isolate organic cloud growth from investment gains. But from a capital markets perspective, it also illustrates why these raises haven't stressed liquidity. The money stays in the system. It just moves from one pocket to another within the same hyperscaler complex.
Late Cycle Signals and the Capex Question
The combined 2026 capital expenditure run rate for Google, Amazon, Microsoft, and Meta now exceeds $700 billion annualized. That's larger than the U.S. federal Medicare budget. Q1 alone saw the big four spend $130.65 billion on capex, more than three times the inflation-adjusted cost of the Manhattan Project.
Historically, capex surges of this magnitude have arrived late in economic cycles, often presaging either a productivity boom or a painful capital destruction phase. The 1998-2000 telecom infrastructure buildout is the closest analog. Back then, excess fiber capacity took nearly a decade to absorb, and many companies that financed the buildout went bankrupt. The difference today is that the spending is concentrated among companies with fortress balance sheets and recurring revenue streams that can absorb writedowns.
The counterargument is that AI revenue growth may justify the spending. Anthropic's annualized revenue reportedly jumped from $1 billion at the end of 2024 to $30 billion by early 2026. That's 30x growth in roughly 15 months. If the revenue trajectory holds, the capex looks aggressive but rational. If growth plateaus, the hyperscalers will face tough questions about stranded capacity.
What Credit Spreads Are Telling Us
Investment-grade credit spreads have barely budged through the announcement window. That's the market's way of saying it doesn't see immediate stress. IG spreads typically widen when large capital commitments threaten to crowd out other borrowers or signal balance sheet risk. Neither is happening here.
Part of the explanation is the staged nature of the commitments. Google's $30 billion contingent tranche only deploys if Anthropic hits performance targets. Amazon's structure is similar. The upfront cash commitments are large but not enormous relative to the companies' free cash flow generation. Alphabet generated over $70 billion in operating cash flow last fiscal year. A $10 billion investment, while significant, doesn't threaten the balance sheet.
Another part is market structure. With yields on 10-year Treasuries hovering near their recent range, institutional investors remain hungry for duration and spread. The hyperscalers could tap bond markets tomorrow if they needed to fund the investments with debt, and they'd find eager buyers.
The IPO Overhang
Anthropic is reportedly considering a public offering as soon as October. At a $350 billion private valuation, the company would debut as one of the largest IPOs in history, comparable in scale to Saudi Aramco's 2019 listing. The tender offer Anthropic made to employees earlier this year fell flat. Insiders refused to sell, expecting the IPO price to exceed the $350 billion mark.
If the IPO proceeds, it will test public market appetite for AI exposure at nosebleed valuations. The company trades at under 12 times annualized revenue based on the latest figures. That multiple would be reasonable for a mature enterprise software company growing 15% annually. For a company growing 3x per quarter, it looks almost cheap. The question is whether that growth rate is sustainable once the infrastructure deals stabilize and competition intensifies.
Public markets have been selective about AI exposure. Nvidia trades at a premium to its own historic multiples. Pure-play AI names have struggled. Anthropic's debut will reveal whether the private market froth translates to public demand or whether investors balk at the implied profitability timeline.
What to Watch Over the Next Month
The setup is stable for now, but a few things could shift the picture. First, watch credit spreads in the tech sector. If IG spreads begin widening despite strong earnings, it may signal that investors are repricing the sustainability of capex commitments. Second, monitor the Fed's posture on financial conditions. A surprise hawkish turn could tighten liquidity just as the hyperscalers need to absorb these commitments.
Third, and most importantly, track whether Anthropic files its S-1 registration. An October IPO would need to start the SEC process within weeks. A delay might indicate that the company or its underwriters see softness in demand. Conversely, an aggressive timeline would signal confidence that public markets are ready to absorb another massive AI valuation.
The buses have arrived. Whether the market has enough seats for everyone remains the open question. You can track institutional flows and sector rotation in real time on the [Sector Rotation dashboard](/sector) as these capital commitments work through the system.
For informational purposes only. Not investment advice. Published Tuesday, June 2, 2026.