Whiteprint
Equity Research
Equity Research

Nvidia Valuation Model

A deep-dive valuation analysis on Nvidia, examining whether current multiples are justified by the AI infrastructure buildout thesis.

Investment Summary

Nvidia (NVDA) remains the defining equity story of the AI cycle. With data center revenue now exceeding $40B per quarter, the company has established an unassailable lead in GPU compute for AI training and inference workloads. The question is not whether Nvidia is a great business -- it is -- but whether the stock at current levels adequately compensates for the risks ahead.

Business Overview

Nvidia's revenue concentration in data center GPUs (now ~85% of total revenue) is both a strength and a vulnerability. The H100/H200 product cycle delivered unprecedented pricing power, and the Blackwell architecture extends this advantage. Key customers include hyperscalers (Microsoft, Google, Amazon, Meta) and sovereign AI programs.

Valuation Framework

We employ a three-stage DCF model:

Stage 1 (2026-2028): High growth phase. Revenue grows at 25-35% annually as Blackwell ramps and inference demand accelerates. Operating margins remain above 60%.

Stage 2 (2029-2032): Normalization. Growth decelerates to 10-15% as the market matures and competition from AMD, custom silicon (Google TPU, Amazon Trainium), and new entrants intensifies.

Stage 3 (2033+): Terminal. 3-4% perpetuity growth rate with margins compressing toward 45-50% as the compute market commoditizes.

Key Assumptions

Metric2026E2027E2028E
Revenue ($B)175215260
Gross Margin74%72%70%
Operating Margin62%60%58%
FCF Yield2.8%3.2%3.8%

Risk Factors

  1. Customer Concentration -- The top 4 cloud hyperscalers represent over 50% of data center revenue. Any shift toward custom silicon reduces the TAM.
  1. Cyclicality -- Despite the structural growth narrative, GPU spending is capex-driven and inherently cyclical. A capex pullback from cloud providers would hit Nvidia disproportionately.
  1. Export Controls -- Restrictions on sales to China have already reduced the addressable market. Further geopolitical tensions could expand these limitations.
  1. Valuation Multiple Compression -- At 35x forward P/E, any deceleration in growth expectations will be met with severe multiple compression.

Price Target

Our base-case 12-month price target implies a mid-single-digit upside from current levels. The stock is fairly valued at this point -- the AI thesis is largely priced in. We rate Nvidia as Hold with a focus on monitoring the Blackwell ramp trajectory and hyperscaler capex guidance.

Conclusion

Nvidia is an exceptional business that is approximately fairly valued. The risk-reward is no longer asymmetric to the upside. We prefer to wait for either a better entry point or evidence that the next product cycle (Rubin) extends the growth runway beyond current estimates.