Is NVDA Overvalued in 2026? A DCF Analysis

NVIDIA ($NVDA) has become the defining stock of the AI era. From a market cap of ~$300B in early 2023 to briefly crossing $3 trillion in 2024, it's one of the fastest wealth-creation events in market history. But at current prices, the obvious question for any serious investor is: how much of the future is already priced in?

In this analysis, we run a discounted cash flow (DCF) model on NVDA using current fundamentals to estimate its intrinsic value — and tell you plainly whether the current price offers a margin of safety.

Note: This is not financial advice. Always verify figures against the latest SEC filings before making investment decisions.

NVDA by the Numbers

Before modeling, let's anchor the inputs. As of early 2026:

| Metric | Value | |--------|-------| | Revenue (TTM) | ~$130B | | Net Income (TTM) | ~$73B | | EPS (diluted, TTM) | ~$2.94 | | Free Cash Flow (TTM) | ~$60B | | Revenue Growth (YoY) | ~114% | | Operating Margin | ~62% | | Data Center Revenue Mix | ~88% of total |

What jumps out immediately: NVDA is not a typical semiconductor company. It's generating software-company margins on hardware revenue. The question for valuation is simply: how long can this last, and what's a fair price for that durability?

The DCF Analysis: What Is NVDA Actually Worth?

A DCF requires three key inputs. Here's how we set them for NVDA.

Input 1: Growth Rate

NVDA's recent growth rate (114% YoY) is not a reasonable base case for a 10-year model — nothing sustains that. The real question is where growth decelerates to.

We model three scenarios:

Bear case: Growth decelerates sharply as AI capex normalizes and competition (AMD, Intel, custom silicon from Google/Amazon/Microsoft) bites. EPS grows at 12% annually over 10 years.
Base case: NVDA maintains AI infrastructure dominance through the decade. EPS grows at 20% annually — aggressive but defensible given the software/CUDA moat.
Bull case: Sovereign AI, robotics, and autonomous vehicles add a second growth wave. EPS grows at 28% annually.

Input 2: Discount Rate

For a company with NVDA's risk profile — cyclical underlying hardware demand, geopolitical exposure (TSMC concentration, China export controls), significant customer concentration — we use 10% as the required return. Some analysts use 9% (citing NVDA's durable moat); more skeptical investors use 11–12%.

Input 3: Terminal Growth Rate

3% — long-run GDP growth. Standard assumption for a large-cap mature business.

The Results

Using a 10-year DCF, discounting projected earnings back to present value, then adding terminal value:

| Scenario | Growth Rate | Implied Fair Value | |----------|-------------|-------------------| | Bear | 12%/yr | ~$65–80 | | Base | 20%/yr | ~$130–160 | | Bull | 28%/yr | ~$250–320 |

All figures per share (diluted). Ranges reflect sensitivity to terminal growth assumptions.

What the Market Is Currently Pricing In

This is the more useful question. Rather than asking "is it cheap?" — ask: at the current price, what growth rate is the market pricing in?

Working backwards from the current share price (~$115–125 range as of this writing), the market is implying roughly 22–25% annual EPS growth for the next decade with a 10% discount rate.

That's aggressive. It's not insane given NVDA's position — but it leaves essentially no margin of safety. You're paying for the base case and getting no discount for execution risk, competitive disruption, or macro deterioration.

Value investors typically want a 20–30% discount to intrinsic value before pulling the trigger. At a base-case fair value of ~$145, a 25% margin of safety means a target entry price around $108 or below.

The Bear Case for NVDA

Why the current price might deserve a lower valuation:

Competition is accelerating. AMD's MI300X is gaining data center traction. Google's TPUs, Amazon's Trainium, and Microsoft's custom silicon all reduce hyperscaler dependency on NVDA.
Customer concentration is extreme. Four hyperscalers (Microsoft, Google, Amazon, Meta) account for a disproportionate share of Blackwell revenue. If any one slows capex, it shows up fast.
Export controls are a real risk. US restrictions on H100/H200/Blackwell sales to China have already cost NVDA billions in revenue. Further tightening is plausible.
Inventory cycles happen. Semiconductor cycles are brutal and real. Datacenter capex can slow abruptly.
Valuation math is unforgiving at high multiples. At 35–40x earnings, a miss on growth expectations causes disproportionate price declines.

The Bull Case for NVDA

Why NVDA could be worth significantly more:

CUDA moat is extraordinarily deep. 15+ years of developer ecosystem lock-in. Competitors are building the hardware; they're still catching up on the software.
Sovereign AI is an underappreciated tailwind. Every G20 country is spending on national AI infrastructure. NVDA is the default vendor.
Gross margins at 73%+ on hardware are structurally unusual — they reflect software/ecosystem pricing power, not hardware economics. This can persist longer than skeptics expect.
Robotics and autonomous systems are early innings. If NVDA's Jetson platform captures a fraction of the robotics market, it's a meaningful second growth engine.
Blackwell demand exceeds supply. Constrained supply into 2026 means pricing power is intact near-term.

How to Run Your Own NVDA Valuation

The scenarios above are ours. Yours might differ — especially on growth rate assumptions and your required return. A value investor with a 12% hurdle rate gets a very different number than one using 9%.

The best investment decisions come from running your own numbers.

EvidInvest shows you the fair value estimate for NVDA (and 10,000+ other tickers) instantly — including the implied growth rate the current price is baking in, and your margin of safety at any entry price.

Run a free NVDA valuation on EvidInvest →

Frequently Asked Questions

Is NVDA stock a good buy in 2026?

At current prices (~$115–125), NVDA is priced for a base-to-bull outcome with minimal margin of safety. It's not obviously overvalued — the fundamentals are exceptional — but you're paying for perfection. A pullback toward $100–110 would offer a more comfortable entry for value-oriented investors.

What is NVDA's intrinsic value?

Using a DCF with 20% annual EPS growth (base case) and a 10% discount rate, we estimate NVDA's intrinsic value at approximately $130–160 per share. This is highly sensitive to growth assumptions — the bull case (28% growth) reaches $250–320. Run your own inputs at EvidInvest.

Is NVIDIA a value trap?

No. A value trap is a cheap stock with deteriorating fundamentals. NVDA is the opposite problem — exceptional fundamentals at a full price. The risk isn't a trap; it's paying too much for a great business and earning mediocre returns.

How do you value a high-growth stock like NVDA?

DCF is the right framework, but the inputs matter more than the formula. For NVDA, the critical assumption is the growth rate 5–10 years out, not the near-term revenue (which is strong). Focus on: how long can AI infrastructure capex grow at this rate? What does normalized EPS look like in 2030? Work backwards from there.

What P/E ratio is NVDA trading at?

At ~$120/share with TTM EPS of ~$2.94, NVDA trades at roughly 40x trailing earnings. On a forward basis (FY2027 estimates ~$4.50–5.00 EPS), it's approximately 24–27x forward earnings — expensive but not absurd for a 20%+ grower with 60%+ operating margins.

Our Take

NVDA is a genuinely great business at a full price. The AI infrastructure buildout is real, the CUDA moat is real, and the margin profile is exceptional. But at current prices, you're not getting a discount — you're paying roughly fair value for the base case.

Our view: watch for a pullback toward $100–110 for a better risk/reward entry. Set a price alert. And stress-test your own assumptions before committing capital — the difference between 20% and 15% growth over 10 years is enormous in present-value terms.

Try EvidInvest free — 14-day trial, no credit card required →