5 Billionaire Money Managers Dumped Palantir in the Second Quarter. This Is the AI Applications Stock They're Buying Hand Over Fist Instead.

13F filings show five billionaire asset managers selling Palantir in the June-ended quarter, while four billionaires continued buying Alphabet. Palantir’s valuation remains extremely stretched (P/S ~73, down from >100 at the start of the year), implying sustainability risk. Meanwhile, Alphabet’s AI-driven Google Cloud strength stands out (Q2 revenue +24% Y/Y to $119.8B; Google Cloud revenue +82% Y/Y to $24.8B; operating margin 34% and Google Cloud operating margin 36%), supporting a more constructive read-through despite the mixed positioning signal across AI stocks.

Analysis

The real signal in the flow data is not simply "AI winners vs losers"; it is which monetization model is attracting durable capital. That favors GOOG/GOOGL because its AI spend is embedded inside two existing profit engines — ads and cloud — so incremental AI revenue is more likely to be margin-accretive than speculative. PLTR, by contrast, still depends on the market paying a scarcity premium for future optionality; once a few large holders de-risk, the stock can re-rate quickly because the ownership base is more momentum-sensitive than fundamental.

Near term, the 13F read is mostly a positioning indicator, not a catalyst by itself. The first 1-2 weeks could see noise in both names, but the cleaner 1-3 month path is GOOG if earnings revisions keep creeping up on cloud and ad resilience. PLTR’s risk is an air pocket if the next print fails to show broadening commercial adoption; at 70x+ sales, even a small growth miss can compress multiple faster than the business can grow into it.

The contrarian miss is that the market is treating all AI exposure as one trade, when the balance-sheet and distribution advantage matter more than the narrative. GOOG can self-fund AI capex and defend returns; PLTR needs persistent buy-side sponsorship to justify its valuation. If that sponsorship keeps bleeding, the downside is not linear — it is a crowded-long unwind, which historically is where the biggest underperformance happens.

Key Decisions for Investors

• Long GOOG / short PLTR as a 1-3 month pair trade; size beta-neutral. Thesis: monetized AI + balance-sheet support should outperform a high-multiple crowding unwind. Falsify if PLTR guides materially above current growth expectations or GOOG cloud growth/margins decelerate.
• Initiate GOOG 3-6 month call spread on any market pullback rather than chasing strength. Risk/reward is best if implied vol stays subdued and the market keeps rewarding profitable AI exposure; exit if cloud growth slows meaningfully for two quarters.
• Use PLTR 1-3 month put spreads, not outright shorts, to express valuation compression risk. This keeps risk defined if the company prints another upside surprise; cover if revenue growth re-accelerates and management raises full-year guidance.
• If already long broad AI baskets, rotate marginal capital from PLTR into GOOG/GOOGL and cash-rich platform names. The second-order effect is a widening dispersion trade: cash-generative AI platforms should outperform story stocks as rates volatility returns.