Jim Cramer drops blunt call on Nvidia stock

Nvidia reported a blockbuster Q3 with revenue of $57 billion (up 22% sequentially and 62% YoY), data‑center sales of $51.2 billion (up 25% sequentially and 66% YoY), net income near $31.9 billion, and guidance of roughly $65 billion for the next quarter. Despite those fundamentals, the stock has fallen roughly 4.71% in the past week (about $200 billion in market cap) amid investor fears over competition from Google’s TPUs and Meta’s plans and valuation warnings from Michael Burry; Jim Cramer counters that the selloff is fear‑driven and presents a buying opportunity. Nvidia also defended its buyback accounting (citing ~$91 billion repurchased since 2018 and cash flow roughly 98% of reported net income), while market positioning and analyst attention remain intense given Nvidia’s central role in the AI trade.

Analysis

Market structure: Nvidia remains the dominant supplier for large-scale training (Q3 data‑center sales $51.2B; company guiding ~$65B next quarter), but Google’s TPU push and Meta renting TPUs could reallocate ~10% of Nvidia’s AI‑chip sales by 2026–2027. Winners are vertically integrated cloud players (GOOGL, MSFT) and interconnect/analog suppliers (AVGO) that monetize multi‑cloud demand; pure GPU incumbents (single‑stack AMD, smaller ASIC specialists) face mixed outcomes. The immediate supply/demand picture is tight — backlog jumps (Google cloud backlog +$49B) and constrained HBM/TSMC capacity imply price power remains with incumbent chip vendors for at least 12–18 months.

Risk assessment: Tail risks include (1) regulatory/antitrust action vs dominant GPU ecosystems, (2) a faster‑than‑expected shift to vertically integrated TPUs/ASICs eroding NVDA revenue 10–20% over 2–3 years, and (3) a foundry/HBM bottleneck causing sudden margin compression. Near term (days–weeks) expect 5–15% volatility around headlines; medium (3–9 months) depends on major wins/losses from Google/Meta partnerships; long term (2+ years) hinges on software portability (CUDA lock‑in) and model architecture changes. Hidden dependency: Nvidia’s pricing power depends on software lock‑in and HBM supply, not just die performance.

Trade implications: Tactical trades favor overweighting cloud software/infrastructure (GOOGL, MSFT) and select chip supply chain names (AVGO, AMD) while using hedges on NVDA. Specific option/play: sell a portion of NVDA exposure into any >10% bounce and buy 3–6 month protective puts if NVDA falls >12% in 2 weeks. Position sizing should be conservative: tranche into NVDA (1–3% portfolio) and add on confirmed macro/competition pullbacks; add GOOGL (2–4%) for durable cloud exposure within 1–3 months.

Contrarian angle: The market may be over‑pricing a rapid loss of Nvidia share while under‑pricing persistent demand and buybacks (company cites ~$91B repurchases since 2018 and ~98% cashflow/net income conversion). Historical parallel: Cisco in 2000 took years to justify valuations — possible here too, implying multi‑year patience or option hedges rather than outright shorts. If NVDA drops >15% from today, that’s a tactical accumulation trigger; if Google proves TPU parity for LLM training before 2027, rotate 50–75% of NVDA gains into GOOGL/AVGO within 30 days.

Key Decisions for Investors

• Establish a phased long in NVDA sized 1–3% of portfolio: 0.5–1% now, add another 0.5–2% if NVDA drops an additional 10–15% within 30 days or if next‑quarter guidance misses by >5% (sales < $61.75B).
• Allocate 2–4% long GOOGL (cloud/TPU exposure) within 2–6 weeks as a hedge against NVDA share loss; increase to 4–6% if Google Cloud growth in next two quarters stays >30% YoY or backlog increases by another >$20B.
• Buy 3–6 month protective puts on NVDA equal to 25–50% notional of the long position (choose strikes ~10–15% OTM) or implement a collar: buy puts and sell near‑term calls if seeking to reduce cost; roll as headlines develop.
• Initiate 1–2% long in AVGO and 1% long in AMD as supply‑chain/alternative GPU exposure; take profits if NVDA recovers >15% from current levels or if Google/Meta announce large TPU commitments removing >5% of NVDA’s revenue runway.
• If NVDA downside exceeds 15% from today, convert protective positions into additional long tranches (buy to total 4–6% exposure) and reduce short/overweight positions in NVDA competitors only after concrete TPU production milestones (Meta 2027 commitment) are publicly signed.