Billionaire Money Managers Have Chosen Their Favorite AI Stock (and It's Not Nvidia or Alphabet)

The article argues that billionaire fund managers have been selling Nvidia and Alphabet while favoring Amazon as the leading AI stock, with Amazon appearing in the top holdings of six prominent managers. It highlights Nvidia's 92% data center sales growth and Alphabet's 63% Google Cloud revenue growth, but says Nvidia faces increasing GPU competition and Alphabet is no longer as cheap at a forward P/E near 26 versus 16 a year ago. Amazon Web Services still controls nearly a third of global cloud infrastructure spending and is presented as the most attractive value opportunity among the major AI platforms.

Analysis

The positioning signal matters more than the headline AI winners/losers narrative: if sophisticated capital is rotating from the pure-play GPU beneficiary into the platform operator, that implies the market is moving from scarcity pricing to monetization durability. That is a subtle but important regime shift—when customers can self-design chips or multi-source accelerators, the upside in semiconductor suppliers becomes more cyclical, while the value migrates to the layer with embedded workloads, switching costs, and incremental margin leverage.

Amazon’s relative appeal is not just valuation; it is optionality with a lower dependency on any single AI architecture. AWS can monetize AI whether inference runs on proprietary silicon, Nvidia hardware, or third-party accelerators, which makes AMZN the cleaner “pick-and-shovel plus distribution” exposure over a 2-3 year horizon. The second-order beneficiary set likely extends to network, storage, and power infrastructure vendors serving cloud capex, while the most vulnerable are vendors whose economics require persistent shortage pricing.

The key contrarian point is that consensus may be overestimating how quickly AI capex translates into durable shareholder returns. A large share of the near-term spend still looks like pre-revenue capacity buildout; if utilization rates lag, the market will start punishing gross-margin compression and depreciation drag rather than celebrating revenue growth. That creates a setup where AMZN and GOOGL can outperform on quality-of-revenue and operating leverage, while NVDA becomes more sensitive to any evidence of order normalization, customer in-sourcing, or slower hyperscaler capex growth over the next 6-12 months.

Key Decisions for Investors

• Long AMZN / short NVDA pair trade over the next 3-6 months: express the view that monetization durability and valuation asymmetry favor the platform owner over the scarcity-priced hardware supplier; target 10-15% relative outperformance if hyperscaler capex growth moderates.
• Add to GOOGL on pullbacks with a 6-12 month horizon: the market is paying up for AI applications, but cash-generation plus distribution remains underappreciated; upside is more measured, but downside is cushioned versus pure infrastructure plays.
• Reduce outright NVDA exposure or hedge with call spreads ahead of the next two earnings cycles: the risk is not demand collapse but multiple compression if customer self-supply and competitive silicon become visible in guidance.
• Basket long AWS-adjacent beneficiaries outside the article’s names — networking, storage, and datacenter power/electrical infrastructure — as the second-order winners from AI capex persists even if semiconductor economics normalize.
• If entering AMZN, prefer staged entry or a 3-6 month call spread rather than common stock: this preserves upside to cloud reacceleration while limiting valuation risk if AI monetization proves slower than the market expects.