I Predicted That Nvidia Would Beat the S&P 500 for the 3rd Consecutive Year in 2025. Here's Why the Streak Can Continue in 2026.

Nvidia has reported a surge in booked orders — previously guided $500 billion in Blackwell and Rubin orders and said at CES that orders have since exceeded that level — and announced six new Rubin chips now in full production with deliveries beginning in H2 calendar 2026. Analysts have materially raised estimates, with consensus EPS now $4.69 for fiscal 2026 and $7.60 for fiscal 2027 (up from $4.29 and $5.76 six months prior), while Nvidia touts Rubin efficiency gains of up to a 90% reduction in inference token cost and a 75% cut in GPUs needed for training. The combination of strong demand, product-driven margin expansion and a ~39x forward P/E supports a constructive investment case, though the piece flags downside risk if hyperscalers eventually pull back on capital expenditures.

Analysis

Market structure: Nvidia (NVDA) is the clear near-term winner — Rubin’s claimed 75% GPU count and up to 90% inference token cost improvements give Nvidia pricing power and justify premium ASPs, benefiting upstream suppliers (TSMC, ASML, LRCX) and rack/network vendors while squeezing GPU-focused rivals (AMD, INTC) and small AI-chip entrants. The $500B+ booked orders signal demand >> supply for next 12–24 months, creating sustained order visibility and likely extended fab allocations; expect elevated implied volatility in NVDA options and widening credit spreads for leveraged smaller chipmakers if revenue misses emerge.

Risk assessment: Tail risks include a hyperscaler capex pause (a 20–30% YoY cut would create a high-impact revenue shock), supply-chain shocks at TSMC or yield setbacks, and export/regulatory actions that could block key markets; these are low-probability but material. Short-term (days–months) risk is event-driven volatility around production ramp commentary and quarterly guides; long-term (quarters–years) risk is customer ROI discipline forcing slower spend conversion of booked orders.

Trade implications: Favor asymmetric exposure — size directional long exposure to NVDA modestly and use structured options to cap downside. Implement relative-value trades (long NVDA/short AMD) to capture share-shift while hedging market beta. Rotate portfolio into semicap/equipment and networking names for 6–18 months while trimming cyclical legacy CPU/PC suppliers.

Contrarian angles: Consensus underestimates timing friction — bookings don’t equal immediate revenue; much of Rubin upside is backloaded into H2 CY2026–FY2027, so forward EPS upgrades may already be priced. Watch for historical parallels (2017 GPU crypto cycle) where rapid demand reversion caused 30–60% drawdowns; set objective triggers (hyperscaler capex growth <+10% YoY or NVDA miss >8% revenue) to materially de-risk positions.

Key Decisions for Investors

• Establish a 2–3% net long position in NVDA now; scale to 5% if NVDA pulls back 10–15% intraday or within 8 weeks; use a hard stop-loss at -12% from entry and target a 12-month upside of +25–40% if Rubin revenue ramps as guided.
• Buy a 9–12 month call spread on NVDA: buy ATM calls and sell 25–30% OTM calls (size = ~1% portfolio) to capture upside from Rubin adoption while financing premium; close if NVDA rises >35% or if quarterly guide reduces expected Rubin shipment cadence.
• Execute a pair trade: go long NVDA (1.0x notional) and short AMD (0.7–1.0x notional) to express expected margin and share gains from Rubin; rebalance or close if the pair diverges more than 15% in 60 days or if AMD announces a competing win with demonstrable cost parity.
• Hedge tail risk: if NVDA position >3% of portfolio, buy 6-month puts 20% OTM sized to cover ~50% of position; additionally, set market alerts to reduce NVDA exposure by 50% within 30 days if aggregate hyperscaler capex guidance (Amazon/MSFT/Google) is cut >10% QoQ.