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

Billionaire fund managers have been net sellers of Nvidia and Alphabet while making Amazon their preferred AI stock, with Amazon appearing in the top-four holdings of six prominent managers. The article argues Amazon offers a better value proposition, citing AWS's roughly one-third share of global cloud infrastructure spending and projected cash flow growth from $12.89/share in 2025 to $27.66/share by 2028. The piece is constructive on Amazon but highlights valuation concerns for Alphabet and competition risks for Nvidia.

Analysis

The positioning signal is more interesting than the headline fundamentals: sophisticated capital is clustering around the most monetizable layer of the AI stack while de-emphasizing the pure compute winner. That implies the market may be transitioning from a supply-constrained phase, where GPU scarcity drove excess returns, into a more normalized capex cycle where system integrators and platform owners capture a larger share of incremental AI economics.

That shift creates a second-order beneficiary set beyond the obvious trio. If large customers continue internal silicon development, NVIDIA’s pricing power can compress before unit demand does, which means the equity can still look operationally strong while multiple compression begins to outrun earnings growth. On the other side, the real leverage in AI may be migrating to the cloud layer, where incremental workload monetization expands gross profit dollars faster than revenue, making Amazon’s operating model more resilient than Alphabet’s despite slower top-line growth.

The key contrarian miss is valuation dispersion versus durability. The consensus is still rewarding “AI exposure” indiscriminately, but the highest-quality trade may be the one with the cleanest path to cash flow compounding at a still-reasonable multiple. Amazon’s setup is stronger because the market can re-rate it on margin expansion alone, whereas NVIDIA and Alphabet now require either sustained hyper-growth or a better-than-expected deceleration path to avoid multiple give-back.

Risk is mostly over a 3-12 month horizon: a broad AI spending pause, faster-than-expected in-house accelerator adoption, or evidence that cloud AI monetization is being subsidized rather than accretive. The asymmetric setup favors owning the company that can absorb a normalization in AI infrastructure spend and still compound free cash flow, while being more selective on names that already discount perfection.

Key Decisions for Investors

• Overweight AMZN vs. basket: initiate a long AMZN / short equal-dollar NVDA + GOOG pair trade for 6-12 months; thesis is cash-flow compounding and multiple resilience at the cloud layer versus normalization risk in compute and ad-driven AI monetization.
• Use NVDA as a tactical trading name, not a core hold: sell upside into strength or buy put spreads 3-6 months out if data-center growth remains hot but supplier/customer capex commentary suggests internal chip substitution; risk/reward favors downside on any demand-scarcity normalization.
• Keep GOOG as a quality compounder but avoid chasing here: wait for a valuation reset or deploy via call spreads only on post-earnings volatility; current setup looks more fully valued than AMZN with less direct AI cash-flow leverage.
• If seeking a cleaner AI beta expression, rotate into AMZN on 1-3 month weakness near support rather than buying breakouts; reward is multiple expansion plus operating leverage, while downside is partially cushioned by diversified cash generation.