Palantir Billionaire Peter Thiel Just Bought Shares of This AI Stock That’s Climbed 200,000% Since Its IPO -- And It’s His Top Position.

Thiel’s Thiel Macro added to AI-adjacent exposure by buying 495,000 shares of Amazon, bringing AMZN to 28% of his portfolio after investing $418M across eight stocks. Amazon reported its chips business revenue rose to an annual run rate of $25B+ (and AI revenue to $25B+), supporting the view that AWS can accelerate AI-driven demand. The article also cites AMZN trading at ~20x forward earnings, framing the setup as “dirt cheap” versus its earnings track record, though the move is portfolio-specific rather than market-wide.

Analysis

This is less a "Thiel loves AI" signal than a capital-allocation clue: the best AI economics may accrue to the infrastructure owner that can sell compute, chips, and workflow automation off one balance sheet. AMZN is the cleanest way to own both the demand and the margin lever, because AI spend there can show up as higher AWS mix and lower fulfillment cost-to-serve. If the AI revenue run-rate is durable, the market should eventually stop valuing AMZN like a retail compounder and start giving it a higher software/cloud multiple on the earnings stream that matters most.

The second-order winner is VST, not the regulated utilities basket. Merchant power can reprice quickly when data-center demand tightens local supply, while AEP/CMS/DTE/FE mostly need regulatory approval to translate load growth into earnings, which is a 12-24 month process at best. That creates a likely mismatch: headline-driven inflows into utilities now, but better medium-term economics in power producers with market pricing power and less regulatory lag.

Risk is that the AI capex cycle outruns monetization. If AWS growth slows or capex stays elevated, AMZN can become a margin-story disappointment even if absolute revenue keeps rising. For the utility names, the near-term catalyst is not AI enthusiasm but signed datacenter PPAs, interconnect wins, and rate-case outcomes; without those, the trade can fade once investors realize the load story is more aspirational than earned cash flow.

Contrarian view: consensus may be overpaying for "AI exposure" in bond-proxy utilities and underappreciating that the real bottleneck is power scarcity, not regulated load growth. That favors AMZN and VST more than the whole basket, and it argues for buying the actual cash-flow beneficiaries rather than the headline proxies.

Key Decisions for Investors

• Long AMZN on pullbacks over the next 1-3 months; thesis is AWS AI monetization plus fulfillment automation, with 6-12 month upside if cloud margins hold. Falsify if AWS growth decelerates or capex rises faster than AI revenue.
• Pair trade: long VST / short equal-weight basket of AEP, CMS, DTE, and FE for 6-12 months. View is merchant power monetizes AI load faster than regulated utilities; cover if utilities begin signing material datacenter PPAs or rate cases reprice earnings upward.
• Underweight or fade rallies in AEP, CMS, DTE, and FE until there is independently verifiable evidence of incremental earnings from AI-driven load. These are duration proxies; the AI narrative alone is not enough to justify multiple expansion.
• Set an alert on AMZN next earnings: if AI-related revenue remains above the current $25B run-rate and operating margin inflects, add to longs; if not, trim because the market may already be discounting the story.