The Last Time Nvidia Stock Was This Cheap, It Nearly Doubled in 6 Months. Can It Repeat?

Nvidia's shares, after tumbling to roughly 24x forward earnings during the April–May 2025 tariff-driven selloff and surging above 40x to deliver an ~81% rally, have retraced to about 25x forward earnings, a valuation the author describes as 'cheap.' Wall Street forecasts revenue growth of ~52% for fiscal 2027 (ending Jan 2027), a modest deceleration from an expected 63% in fiscal 2026, while management projects global data-center capex expanding to $3–4 trillion annually by 2030 and hyperscalers have flagged record capex into 2026. Given Nvidia's dominant GPU position in AI computing and sustained hyperscaler demand, the piece argues the stock is positioned to materially outperform and could potentially double within two years, though competition and geopolitical/tariff risks are noted.

Analysis

Market structure: Nvidia (NVDA) is the direct beneficiary of a multi-year data‑center capex cycle; hyperscalers (AMZN, MSFT, GOOGL), foundries (TSM), and ASML capture secondary upside via volume and node demand, while legacy CPU vendors (INTC) and smaller GPU rivals will face margin pressure. The shift preserves Nvidia’s pricing power so long as foundry capacity (TSMC) remains constrained and demand growth >40% YoY; a 50%+ CAGR in AI rack demand would keep gross margins elevated into 2027.

Risk assessment: Principal tail risks are renewed China export controls, a hyperscaler capex pause (>15% cut vs plans), or a product execution failure (Rubin delay) that could halve consensus revenue growth; these are low probability but high impact within 3–12 months. Hidden dependencies include TSMC wafer allocation and power/data‑center build bottlenecks; monitor TSMC lead times and hyperscaler capex release cadence as immediate indicators.

Trade implications: Favor asymmetric long exposure to NVDA sized 1–3% of AUM using equity + defined‑risk options to avoid IV premiums; overweight TSMC/ASML and top cloud spenders to play the value chain. Consider a relative‑value pair (long NVDA, short INTC) to neutralize market beta and use profit targets (double or forward P/E >40) and hard stops (earnings miss >5% rev or China revenue loss >20%).

Contrarian angles: Consensus underprices geopolitical concentration and inventory cycles — if hyperscalers vertically integrate or diversify accelerators, NVDA could lose upside faster than current models assume. Options IV is elevated; prefer spreads to outright long calls. Historical analogue: 2016–18 GPU cycles show fast upside but sharp corrections on capex pauses, so size positions accordingly.

Key Decisions for Investors

• Establish a 2% long position in NVDA (equity) if forward P/E ≤26 or share price declines another -5% from current levels; target hold through Jan 2027, trim half if forward P/E >40 or position value doubles, stop-loss if company cuts FY guidance >5%.
• Buy a defined‑risk bullish options spread sized 1% AUM: 12–24 month NVDA LEAP call spread (buy ATM+10% expiry Jan 2028, sell 30% OTM) to capture upside to 2027 while capping premium paid; unwind if Rubin architecture misses ship date by >3 months or if hyperscaler capex guidance drops >10%.
• Implement a pair trade: long NVDA (1–2% AUM) and short INTC (0.5–1% AUM) to hedge market beta and express AI compute share gains; rebalance if NVDA/INTC relative performance diverges >25% or following quarterly share commentary changes.
• Overweight supply‑chain beneficiaries: add 1–2% positions in TSM (TSM) and ASML (ASML) to play sustained wafer and lithography demand; reduce cyclical consumer semiconductor exposure by 1–2% and avoid outright long lone GPU peers without proof of volume wins in China within 6 months.