Cloud Earnings, One Hedge Fund, One Month: The Real Shape of AI's Leverage

The following Whitepaper, part of our Technology & AI series, covers a single, connected story that played out across public and private markets in July 2026. Record cloud earnings, a hedge fund’s forced liquidation, and a record rotation of capital into AI infrastructure all happened within the same month and they’re not three separate stories. They are the same capital market repricing AI exposure by leverage, not by theme. This report breaks down what actually happened and what it signals for the rest of 2026.

---

Table of Content

• Executive Summary
• Cloud Earnings Hide Real Leverage
• A Hedge Fund’s Leverage Reckoning
• LP’s Same Bet, Different Structure
• What This Means Going Forward
• Key Takeaways
• FAQ

Executive Summary

Second-quarter 2026 cloud earnings looked extraordinary: AWS grew 36.7% year-over-year, its fastest pace in 18 quarters, while Azure grew 43%. But Amazon’s trailing free cash flow swung to a $7.6 billion outflow, and $53.4 billion of its reported net income came from a non-operating markup on its Anthropic stake, not operations. Days later, Situational Awareness, a hedge fund running roughly four times leverage on AI infrastructure bets, lost 67% of its value in a single month and was forced to sell its public portfolio to Citadel after margin calls from three prime brokers. Meanwhile, real assets fundraising, much of it tied to AI infrastructure, hit a record $206.6 billion in 2025. Capital isn’t leaving the AI trade — it’s leaving the leveraged version of it for one built to survive a drawdown.

Cloud Earnings Hide Real Leverage

Second-quarter 2026 cloud results were genuinely strong on the surface. AWS revenue grew 36.7% year-over-year to $42.2 billion, its fastest growth rate in 18 quarters, and Amazon’s total quarterly revenue crossed $200 billion for the first time in company history, according to Amazon’s Q2 2026 earnings coverage.

Behind the headline, the picture is more complicated. Amazon raised its full-year 2026 capital expenditure guidance to approximately $220 billion, and trailing-twelve-month free cash flow swung to a $7.6 billion outflow from an $18.2 billion inflow a year earlier, driven by a $66.1 billion increase in AI-related infrastructure spending. Reported net income of $62.6 billion included $53.4 billion of non-operating pre-tax income, primarily a markup on Amazon’s investment stake in Anthropic, not revenue generated by operating the business.

The full report explains why this distinction, growth funded by capex and reported through investment gains versus durable operating profit, matters more than the headline beat.

A Hedge Fund’s Leverage Reckoning

The clearest evidence that AI-related leverage had grown excessive arrived from a hedge fund built entirely around the trade. Situational Awareness, founded by a former OpenAI researcher, returned 439% in the first half of 2026 running roughly four times leverage on concentrated AI infrastructure and semiconductor positions, according to reporting on the fund’s collapse.

In July, those positions fell 35% to 47%, and the fund’s total value collapsed 67% in a single month. Three prime brokers, including Goldman Sachs and JPMorgan, issued margin calls the fund could not meet, forcing a sale of its entire public equities portfolio to Citadel. Over the same month, the Philadelphia Semiconductor Index fell 28.6% from its June peak, and a closely watched momentum-technology index dropped 53.5%.

The full report identifies why this was a leverage failure, not a thesis failure, and what that distinction means for other AI-exposed funds still carrying elevated leverage.

LP’s Same Bet, Different Structure

While hedge funds absorbed leveraged losses, institutional allocators were making a related but different move in private markets. Real assets fundraising, spanning infrastructure, energy, and data centers, hit a record $206.6 billion in 2025, pushing its share of total private capital fundraising to 16.4%, nearly double the year before, according to PitchBook’s analysis of the 2025 fundraising shift.

Infrastructure managers closed five funds above $10 billion each in 2025, on par with private equity’s own megafund count, led by a $25.2 billion flagship fund. Unlike leveraged public equity positions, infrastructure investors typically hold long-duration, contracted lease agreements with hyperscalers, generating predictable cash flow regardless of short-term AI stock sentiment.

Firms evaluating this shift for portfolio positioning can see how RCK Analytics’ Technology & AI research supports ongoing analysis of AI-linked capital flows across public and private markets.

The full report explains why this is a change in exposure structure, not a retreat from AI.

What This Means Going Forward

Taken together, these three events point to one dynamic: capital is not leaving artificial intelligence, it is migrating away from leveraged, directional exposure and toward structures built to absorb volatility.

As one senior portfolio manager at a major Wall Street investment firm put it regarding the broader AI trade, the concern isn’t that AI beneficiaries are overpriced, but that positioning around them has become dangerously crowded, a view echoed by Morgan Stanley’s Andrew Slimmon in recent commentary on the tech selloff.

The full report outlines the specific signals, including Q3 2026 free cash flow trends and further potential deleveraging events, that will determine whether this remains a leverage story or becomes something bigger.

Key Takeaways

A hedge fund’s 67% July collapse and cloud earnings propped up by investment gains reveal the same story: capital isn’t leaving AI, it’s fleeing leveraged, directional exposure for contracted, asset-backed structures.

FAQ

Why did AWS and Azure post such strong growth in Q2 2026?

AWS grew 36.7% year-over-year to $42.2 billion, its fastest pace in 18 quarters, while Microsoft’s Azure grew 43%, driven by accelerating enterprise AI infrastructure demand, according to Amazon’s Q2 2026 earnings results. However, this growth required Amazon to raise its 2026 capital expenditure guidance to approximately $220 billion, pushing trailing free cash flow into a $7.6 billion outflow a dynamic we first flagged in our Q2 2026 Earnings Report.

What caused the Situational Awareness hedge fund collapse?

The fund ran approximately four times leverage on concentrated AI infrastructure and semiconductor positions. When those positions fell 35% to 47% in July 2026, the leverage turned the decline into a 67% collapse in fund value, triggering margin calls from three prime brokers and a forced sale of its public portfolio to Citadel, as detailed in reporting on the fund’s collapse. It’s the same “scale doesn’t guarantee post-event stability” pattern we covered in our IPO Pipeline & M&A Report, applied to a leveraged hedge fund instead of a newly public stock.

Is money leaving the AI trade after the July 2026 selloff?

Not based on the data. Real assets fundraising, much of it tied to AI infrastructure like data centers, hit a record $206.6 billion in 2025, according to PitchBook’s analysis of the 2025 fundraising shift. Capital appears to be moving away from leveraged, directional AI equity exposure and toward contracted, asset-backed structures tied to the same underlying buildout, echoing the cloud backlog concentration we flagged in our IPO Pipeline & M&A Report.

How much of Amazon’s Q2 2026 profit came from operations versus investment gains?

Of Amazon’s $62.6 billion net income, $53.4 billion was non-operating pre-tax income, primarily a markup on its investment stake in Anthropic, according to analysis of Amazon’s Q2 2026 results. Operating income, a cleaner measure of the underlying business, rose 43% to $27.5 billion the same “beat rate vs. real signal” distinction we raised in our Q2 2026 Earnings Report.