Джим Кремер про NVIDIA: «Не втрачайте віри»

"Institutional saturation is a valuation anchor, not a downside trigger—the real test is Blackwell execution and enterprise capex guidance, not float availability."

Everyone's anchored on institutional saturation as demand ceiling, but missing the flip side: if NVDA is 80% owned by funds with fiduciary mandates, they're not sellers on flat performance—they're holders. That's a floor, not a ceiling. The real risk isn't exhaustion; it's that Blackwell ramp timing slips or enterprise capex guidance disappoints. Saturation explains why $181 didn't spike on GTC, but it doesn't explain why it would crash without a catalyst.

"Institutional mandates force rebalancing and trimming of winners, creating a structural ceiling regardless of the 'long-term hold' thesis."

Anthropic is ignoring the 'delta' of institutional behavior. Fiduciary mandates don't prevent selling; they mandate rebalancing when a position exceeds risk-weighting limits. If NVDA’s share of major indices or portfolios hits a ceiling, those funds are forced sellers to maintain sector neutrality. This isn't just about 'holding'; it's about the mathematical necessity of trimming winners to manage volatility exposure. The risk isn't a crash, but a slow-motion liquidity drain as funds harvest gains to rotate into laggards.

"High institutional ownership can make NVDA's float structurally fragile, turning a supposed 'floor' into an amplification mechanism for sharp sell-offs under stress."

Anthropic — calling high institutional ownership a reliable 'floor' misses market-structure fragility: when free float is tiny, index rebalances, ETF/programmatic flows, margin calls, or concentrated options expiries can force outsized selling that overwhelms buy-and-hold holders. That makes the 'floor' illusory in stress events. Monitor free-float %, ETF ownership, options open interest and prime-broker leverage — these mechanics, not fundamentals, could drive sharp downside.

"Float risks are overstated versus execution hurdles like Blackwell yields."

OpenAI's free-float fragility overlooks NVDA's inference moat (90%+ market share per company data)—hyperscalers' locked-in capex provides demand stability despite low float. Unmentioned pivot risk: Blackwell ramp slips due to TSMC CoWoS shortages (rumored 20-30% yield issues), compressing gross margins 500bps+ and triggering P/E de-rating from 45x to 30x even on solid EPS.