An independent two-stage DCF analysis by a frontier AI model.
NVIDIA is the undisputed winner of the AI infrastructure buildout. H100/H200/B100 GPUs power virtually every frontier AI model, and CUDA's software moat makes switching costs enormous. Revenue grew 120%+ in FY2025 as hyperscalers raced to build AI capacity.
But semiconductor cycles are real. My 20% FCF growth rate — while enormous — is actually conservative relative to recent results because I expect normalization. Custom ASICs from Google (TPUs), Amazon (Trainium), and startups like Groq will erode NVIDIA's monopoly pricing by 2027-28. The 12% discount rate reflects this cyclicality and concentration risk in a single product line.
NVIDIA is the undisputed winner of the AI infrastructure buildout. H100/H200/B100 GPUs power virtually every frontier AI model, and CUDA's software moat makes switching costs enormous. Revenue grew 12...
A 12.0% WACC reflects NVIDIA Corporation's risk profile, including sector-specific volatility, competitive dynamics, and macroeconomic sensitivity.
A 4.0% terminal rate assumes NVIDIA Corporation grows roughly in line with nominal GDP into perpetuity, reflecting the law of large numbers for a mature large-cap enterprise.
Intrinsic value per share under varying discount rate and terminal growth rate assumptions.
| WACC ↓ / Terminal → | 3.0% | 3.5% | 4.0% | 4.5% | 5.0% |
|---|---|---|---|---|---|
| 3.0% | $123.77 | $108.30 | $96.27 | $86.64 | $78.76 |
| 3.5% | $133.29 | $115.52 | $101.93 | $91.20 | $82.51 |
| 4.0% | $144.40 | $123.77 | $108.30 | $96.27 | $86.64 |
| 4.5% | $157.53 | $133.29 | $115.52 | $101.93 | $91.20 |
| 5.0% | $173.28 | $144.40 | $123.77 | $108.30 | $96.27 |
■ Undervalued vs current price ■ Overvalued vs current price
NVIDIA is the undisputed leader in AI compute infrastructure, with data center GPU market share exceeding 90%. The CUDA moat and full-stack approach (hardware + software + networking) create genuine competitive advantages. However, the stock prices in perfection at 30x+ forward earnings, and the cyclical nature of semiconductor spending means any demand slowdown would cause outsized downside. Customer concentration (hyperscalers) and emerging competition (AMD MI300, custom ASICs) are real risks.
NVIDIA's competitive momentum is extraordinary — data center revenue has grown 100%+ YoY as AI training and inference demand explodes. No other company is growing this fast at this scale.
NVIDIA's moat is primarily CUDA — a software ecosystem that makes switching GPUs prohibitively difficult for developers. But this moat is narrower than it appears: it doesn't apply to inference workloads, and hyperscalers are actively building alternatives.
NVIDIA is the consensus AI pick, which means expectations are priced to perfection. Any miss — even a slight deceleration — would cause significant multiple compression. The bull case requires sustained 40%+ growth for years.
Hypergrowth phases don't last. Opus projects normalization as the initial AI infrastructure buildout matures, competition from custom ASICs intensifies, and hyperscaler capex budgets face scrutiny from investors demanding AI ROI.
Custom silicon. Google's TPU v6, Amazon's Trainium 3, and Microsoft's Maia chips could capture 30-40% of AI training/inference workloads by 2028, breaking NVIDIA's near-monopoly pricing power.
NVIDIA has a genuine technological moat via CUDA, but the current valuation assumes sustained dominance. History shows semiconductor leadership is cyclical. This is not financial advice.
Disclaimer: The numbers presented on this page are for educational and entertainment purposes only. They are the result of a deterministic mathematical model fed with assumptions generated by an Artificial Intelligence (Gemini 3.1). This does not constitute investment advice. Always conduct your own due diligence before investing in the stock market.