Nvidia FY25: Is the Stock Expensive? Not If It Grows 21% Annually for the Next Decade

May 6, 2025 (8mo ago)

This is the billion-dollar question post-FY25 earnings. Revenue skyrocketed from 27BinFY23to27B in FY23 to 130.5B in FY25—a miracle in business history. But investing is about the future. Facing competition from AMD and a potential restructuring of compute demand by "reasoning models," can Nvidia win for another decade? Today, we use hardcore valuation logic to quantify this "bet."


1. The Financials: Beyond Growth, It's Dominance

The core story of FY25 isn't just number inflation; it's a quantum leap in profitability.

  • Quality of Earnings: While revenue grew 126%, gross margin surged from 56.9% in FY23 to 75.0%. This confirms Nvidia has absolute pricing power in the AI chip market—it's not selling hardware; it's selling "tickets to the AI era."
  • Blackwell's Debut: The highly anticipated Blackwell chip contributed $11 billion in revenue in Q4 alone. This silenced concerns about yield rates and proved a seamless transition between product generations.
  • The Concentration Risk: Data Center revenue accounts for 88.3% of the total. This is its strongest weapon but also its biggest vulnerability. If global AI Capital Expenditure (Capex) slows down, this ship will face turbulent waters.

2. The Market Shift: Is DeepSeek an Assassin or an Assist?

The emergence of efficient, low-cost models like DeepSeek has split the market:

  • The Bear Case: DeepSeek proved that top-tier models can be trained on older chips (like H800), significantly reducing the need for cutting-edge GPUs (like H100/Blackwell). Nvidia's "shovel-selling" business has peaked.
  • The Bull Case (Jensen's View): Low-cost models will democratize AI applications, triggering massive inference demand. The demand for inference compute will be 10x that of training. DeepSeek isn't an assassin; it's an assist that expands the total addressable market (TAM).

3. Valuation Decode: What Does Reverse DCF Tell Us?

Instead of guessing where the stock price will go, let's calculate what expectations are currently baked into the price.

We use a "Reverse DCF (Discounted Cash Flow)" model. Logic: Given the current stock price, calculate how much Nvidia needs to earn in the future to justify it.

Key Assumptions Breakdown

  1. Current Market Cap: ~$2.85 Trillion.
  2. WACC (Discount Rate): 15%. This is stringent (tech giants usually have 10%-12%), implying the market treats Nvidia as a high-risk asset requiring higher returns.
  3. Exit Multiple: 20x. Assuming that in 10 years, Nvidia matures to a valuation multiple similar to today's Microsoft or Apple.
  4. Base Cash Flow: FY25E Free Cash Flow to Firm (FCFF) estimated at ~$73.5 Billion.

The Verdict: 21% CAGR

The model shows that to justify the current price, Nvidia needs to maintain a Compound Annual Growth Rate (CAGR) of 21% for the next 10 years.

Is This a Big Bet?

  • Optimistically: Over the past 5 years, Nvidia's CAGR was 47%. Decelerating to 21% for the next decade sounds "conservative."
  • Pessimistically: In human history, very few trillion-dollar companies have sustained >20% growth for 10 consecutive years. This requires AI to be not just a trend, but a 20-year industrial revolution like the Internet.

4. Conclusion: Are You Buying "Faith" or "Math"?

The current stock price doesn't price in a "perfect utopia"; it prices in an "excellent but plausible" growth path (21%).

  • If you believe AGI (Artificial General Intelligence) will rewrite software, autonomous driving, and robotics in the next 5-10 years, Nvidia is still cheap.
  • If you think AI is a short-term bubble like the Metaverse, or that AMD/Intel can steal 30% market share, then the "21% growth" assumption collapses, and the stock faces a repricing.

Disclaimer: This article is for informational purposes only and does not constitute investment advice.