Billionaire Tech Mogul Peter Thiel Bought 8 New Stocks in Q2. They All Have a Common Theme.

Thiel Macro resumed buying in Q2, investing $418M across eight new positions, with all holdings tied to the AI-driven power buildout theme. Amazon and utilities (e.g., AEP at ~10% with ~$42.2M; DTE at ~9.6% with ~$40.3M; FirstEnergy at ~9.5% with ~$40M) reflect a thesis that data-center electricity demand will support capital spending at regulated returns. The article flags key risks—AI demand disruption and potential overbuild—but overall frames the trade as a favorable long-duration positioning for incremental power demand.

Analysis

This is less a clean “AI winner” basket than a duration trade on regulated rate base versus merchant scarcity. The market usually underestimates how quickly hyperscaler load can turn into sanctioned capex, which matters because utility equity upside comes from sustained asset growth, not operating leverage. That favors names with transmission bottlenecks and visible project pipelines, while the near-term cash flow hit sits mostly with the AI capex spenders.

Among the listed names, the higher-quality relative winner is AEP if regulators keep approving incremental investment and load assumptions keep ratcheting up; CMS/DTE/FE are similar but more franchise-specific and less obviously linked to regional transmission scarcity. VST is the more convex expression of the theme because merchant pricing benefits if new load tightens reserve margins faster than supply can be built, but it is also the first to give back gains if AI demand softens or gas-linked generation catches up. AMZN is strategically aligned but financially the least attractive immediate expression because its AI buildout is still a free-cash-flow drag before it becomes an earnings tailwind.

The contrarian point: the consensus is right on direction, but possibly wrong on timing and which equities capture it. Most of the value may accrue to grid equipment, interconnectors, and regulated capital plans over 6-18 months, not to every power-adjacent stock in the basket. The thesis is falsified if hyperscaler capex guidance rolls over, state commissions slow rate-base recovery, or forward load forecasts fail to convert into approved projects over the next 1-2 quarters.

Key Decisions for Investors

• Long AEP on a 1-3 month pullback, targeting 6-12 months: best risk-adjusted expression if utility commissions continue to approve data-center-driven capex; thesis breaks if load growth or capex guidance is revised down.
• Prefer VST over the regulated utility basket for a 3-6 month relative-value trade: long VST vs short XLU or a utility-heavy basket to capture scarcity pricing and reserve-margin tightening, with higher upside but faster drawdown risk if AI demand disappoints.
• Avoid chasing AMZN as the primary power trade; if owning it, keep size small and pair it against a utility long because the market has already priced most of the AI-infrastructure narrative while cash flow remains pressured near term.
• Treat XE as a 2-5 year watchlist, not a near-term trade: only add on financing or utility-offtake milestones; otherwise the binary execution risk dominates the optionality.