Joel Greenblatt Made These 4 AI Stocks His Top Positions. Here's Why They're Worth a Closer Look

Joel Greenblatt’s Gotham Asset Management holds NVIDIA, Apple, Amazon, and Snowflake as AI infrastructure bets, with recent increases in each position. NVIDIA posted $68.1B in Q4 FY2026 revenue, up 73% year over year, while Apple generated $143.8B in quarterly revenue and Amazon’s AWS growth reaccelerated to 24%; Snowflake also grew product revenue 30%. The article is constructive on the group’s fundamentals and valuation, but it is primarily an investing thesis piece rather than new market-moving news.

Analysis

The common thread is not “AI beneficiaries” but balance-sheet translation: each name sits at a different point on the capex-to-cash-flow chain, which matters because the market is starting to discriminate between spenders and monetizers. NVDA remains the cleanest toll-collector, but the second-order winner is likely the networking and systems ecosystem, where demand can compound even if unit growth in accelerators normalizes. That makes the trade less about owning the headline GPU winner and more about staying long the highest attach-rate parts of the stack.

A bigger hidden beneficiary is the enterprise software layer, where adoption can lag infrastructure by several quarters and then re-rate quickly once customers standardize on data platforms. SNOW’s setup is interesting precisely because it is the least consensus-friendly: if AI workloads migrate from experimentation to production, the market may underappreciate the durability of data gravity and switching costs. The risk is not weak fundamentals but a prolonged “proof period” where investors demand visible AI monetization before assigning a higher multiple.

AAPL and AMZN are different expressions of the same theme: one monetizes through installed-base upgrade cycles, the other through infrastructure and ad monetization leverage. Apple’s low capex intensity gives it an unusually powerful free-cash-flow conversion advantage if AI drives even modest device refreshes, while Amazon’s heavy capex creates near-term margin compression that can mask long-duration earnings power. The contrarian point is that consensus may be too focused on absolute AI spend and not enough on who can convert that spend into recurring cash generation fastest over the next 12-24 months.

Key Decisions for Investors

• Long NVDA vs short a basket of lower-quality AI capex beneficiaries for 3-6 months; thesis is that winners will shift toward true toll-collectors as the market re-rates cash conversion over narrative.
• Buy AMZN on any post-capex pullback and express with 6-12 month call spreads; upside comes from AWS reacceleration plus ad leverage, while the main risk is temporary margin disappointment.
• Initiate a tactical long SNOW position only on weakness, with a 6-9 month horizon; asymmetric upside if AI data workloads become productionized, but size small because multiple support is still fragile.
• Prefer AAPL as a defensive AI participation trade over higher-beta software names; use it as a core long if you expect device upgrade cycles to show up over the next 2-4 quarters.
• Hedge the basket with a short in the most crowded AI infrastructure proxies if capex growth data starts decelerating; the key risk event is a downgrade cycle tied to hyperscaler budget normalization.