Palantir Co-Founder Peter Thiel's Hedge Fund Just Made a Big Bet on This AI Stock

Peter Thiel's Thiel Macro held Amazon as 28% of its portfolio at the end of Q2, its largest disclosed equity position, alongside energy-company investments tied to AI power demand. AWS generated 21% of Amazon's Q2 revenue but 60% of operating profit, while its growth rate reportedly accelerated from the low-20% range in 2025 to 37%. The article argues Amazon's planned $220 billion of annual data-center capital expenditures could drive further AWS revenue and profit growth as AI compute capacity comes online.

Analysis

The investable issue is not whether AWS can grow, but whether incremental AI capacity earns above its cost of capital before depreciation and power costs reset the segment's margin profile. AMZN's retail and advertising cash generation can subsidize a longer buildout than most cloud peers, giving AWS latitude to price aggressively for strategic workloads; that is unfavorable for smaller GPU-cloud providers and constrains near-term margin upside at MSFT and GOOGL. The market will reward evidence that capacity utilization is rising faster than the fixed-cost base, not headline capex alone.

Near term, the disclosed fund position is not a catalyst: 13F information is stale and a concentrated macro fund can reverse without notice. The next 1-3 month catalysts are AWS backlog/RPO growth, management commentary on available capacity, and whether quarterly AWS operating margin remains resilient as new facilities are commissioned. The principal risk is a two-stage cash-flow valley: capex is incurred immediately, while revenue ramps and depreciation pressure arrive with a lag; a deceleration in AWS growth or a step-down in consolidated FCF expectations would compress AMZN's AI premium quickly.

The second-order bottleneck is electricity and data-center electrical equipment rather than chips alone. If hyperscaler construction remains durable, CEG and VST benefit from scarce firm-power economics, while VRT and ETN capture higher-value cooling and power-distribution content. Contrarian view: AMZN may already be appropriately valued versus its cloud peers if AI workload pricing deflates as custom silicon and broader GPU supply improve; the upside case requires monetization per unit of capacity, not simply more deployed capacity.

Key Decisions for Investors

• Accumulate AMZN only on post-earnings confirmation that AWS growth and operating margin both hold or improve; use a 6-12 month horizon and size against a 10-12% downside stop if AWS growth decelerates materially or consolidated FCF guidance is cut.
• Prefer a 6-month pair trade long AMZN / short a basket of smaller AI-infrastructure beta via SMH or a targeted GPU-cloud exposure, rather than an outright mega-cap AI long; the thesis is AWS's balance-sheet-funded capacity advantage. Exit if cloud pricing stabilizes and smaller providers demonstrate sustained utilization and positive FCF.
• For the power bottleneck, initiate a staged long CEG or VST over 6-18 months rather than chasing data-center headlines; use power-price and contracted-load disclosures as entry triggers. Key falsifier is a meaningful drop in forward power curves or evidence that new data-center capacity is delayed by interconnection constraints.
• Maintain VRT and ETN on a watch list for orders, backlog conversion, and valuation pullbacks; buy only if bookings support revenue visibility beyond 12 months. The risk/reward deteriorates sharply if hyperscalers defer construction while still spending on internally designed accelerators.