Is Nvidia Overvalued in 2026? NVDA Valuation, AI Growth and the Risks Investors Are Pricing In

Nvidia is expensive by almost every conventional measure. It's also, by some of those same measures, cheaper than it's been in years.

That contradiction is the whole story here. It's why "is Nvidia overvalued" doesn't have a clean yes or no answer. A high price tag alone tells you little. What matters is whether AI revenue, earnings and free cash flow growth still ahead of the company can justify what's already priced in.

Key Takeaways

• Nvidia's trailing P/E of ~28x sits below its own five-year average of ~60x, despite the stock's size and price.
• Revenue growth accelerated from 73% to 85.2% year over year across the last two reported quarters, outpacing the stock's gains.
• Huawei is scaling AI chip output fast but still trails Nvidia's compute by roughly 13x per chip.
• Analyst sentiment remains heavily bullish, 58 of 61 tracked analysts rate NVDA a Buy or Strong Buy.

Where NVDA Sits Right Now

Before getting into the argument, it helps to see the actual numbers side by side, since most of the debate hinges on how you read the same figures differently.

The Current Snapshot

As of late September 2026, Nvidia trades around $226 a share, a market cap near $5.4 trillion, still the largest public company on earth. Trailing P/E sits near 28, forward P/E around 18-20, and PEG comes in near 0.35, well under the 1.0 line often treated as fair value for a growth stock.

Free cash flow yield sits around 2.4%, thin for the price tag but not unusual for a company still reinvesting heavily into capacity.

How That Compares

Here's the surprising part. Nvidia's trailing P/E of 28 sits below its own five-year average of roughly 60x, and under the semiconductor sector average too. The S&P 500's forward P/E runs near 22x for comparison, meaning Nvidia's forward multiple isn't wildly detached from the broader market's either. That's not the picture most people have in their head when they hear "AI bubble."

A stock trading at 211x trailing earnings next to Nvidia's 28x tells you something about where the market has decided to price hope versus proven cash flow.

Earnings Growing Faster Than the Stock

The multiple compression itself is worth visualizing, since it's the clearest evidence for the bull case in one chart.

Revenue grew 73% year over year in fiscal Q4 2026, then accelerated to 85.2% in fiscal Q1 2027. The stock gained about 24% over the same trailing twelve months.

Earnings caught up to, then ran past, the share price. That's the core of the bull argument, and it's a fact pulled directly from the reported numbers, not a talking point.

Bulls vs. Bears

Nvidia doesn't lack for opinions. Two of the most-quoted voices on Wall Street right now sit on opposite sides of the same set of facts, and both have put their positions on the record publicly rather than just in analyst notes.

The Bull Case, in One Analyst's Words

Dan Ives, Global Head of Tech Research at Wedbush and now Partner at Yorkville Ives, has been the loudest bull voice on Nvidia all year. After the company's fiscal Q1 2027 report, he posted directly to X:

He made the same case on Bloomberg TV after Nvidia's fiscal Q4 2026 results, telling the network he was raising his price target to $300:

Speaking separately to CNBC in July, Ives put a number on the demand gap: Nvidia's chip supply is running roughly 12 to 1 against demand, and he places the AI buildout at "the third inning of the AI revolution," meaning, in his read, roughly 15% of the way through the total spending cycle he expects to play out.

The Bear Case, in One Investor's Words

David Bahnsen, CIO of The Bahnsen Group and a regular CNBC contributor, put the bear case in blunt terms on Squawk Box:

— David Bahnsen, CNBC Squawk Box, June 2026

He pointed straight at Nvidia: "Palantir just had the two greatest quarters it's ever going to have and the stock is down 50%. Nvidia has not moved in nine months."

At the time, Nvidia had just posted 85.2% YoY revenue growth and 74% free cash flow growth, yet the stock was up only 6% over nine months. Great numbers, flat stock, that gap is his bubble signal.

What Growth Does $226 Actually Require

Instead of arguing about whether the multiple is too high in the abstract, it's more useful to flip the question and work backward from the price.

The Reverse-Valuation Model

Using trailing EPS of roughly $7.95, here's the earnings growth implied by different terminal multiples three years out. This is a simplified framework, not a full discounted cash flow model, but it's useful for seeing what's actually baked into today's price.

If the market keeps paying growth multiples, the bar is low given the 85% revenue growth already delivered. If the multiple normalizes toward mature semiconductor levels, Nvidia needs to nearly double earnings in three years just to stay flat. 

The Bull Case

Set aside the analyst quotes for a moment and look at what's actually driving the optimism structurally.

Infrastructure Spending Keeps Climbing

Hyperscaler capex keeps climbing rather than flattening. Meta raised its 2026 capex range to $125-145 billion, and Jensen Huang has projected $3-4 trillion in annual AI infrastructure spending industry-wide by 2030.

The CUDA Moat

Nvidia's CUDA software ecosystem keeps switching costs high even as rivals ship competitive hardware on paper. A faster chip on a spec sheet doesn't automatically win a workload if the software stack around it isn't mature.

Beyond GPUs

Data Center Networking revenue climbed 199% in one recent quarter, showing growth isn't just GPU sales anymore. It's an entire systems business now, interconnects, racks, full platforms.

Margins Are Holding

Non-GAAP gross margin has held near 75% even while absorbing higher compensation costs, unusual pricing power at this scale and this size.

The Bear Case

The risks aren't hypothetical either, and several of them are already visible in this quarter's numbers.

Competition Is Catching Up

AMD's data center revenue grew 107% to $6.7 billion in Q2 2026. Broadcom's AI semiconductor revenue jumped 221% to $16.7 billion. Neither rivals Nvidia's scale, but "no real competitor exists" is no longer an accurate description of the field.

Customer Concentration

A handful of hyperscalers account for an outsized share of Nvidia's revenue, and each has an incentive to eventually design more custom silicon of its own.

Policy Risk

Export policy to China has shifted more than once this year, moving real revenue on and off the table regardless of Nvidia's own product roadmap.

Rate Sensitivity

If hyperscaler spending growth decelerates, or rates stay elevated, the multiple the market is willing to pay compresses regardless of how good the underlying business is.

Huawei and the China Wildcard

This is the part of the bear case that's moving fastest right now, and it's specific to this month.

Huawei's Accelerated Roadmap

At Huawei Connect 2026 in Shanghai this September, chairman Eric Xu said Huawei can't produce enough AI hardware to meet domestic demand. He confirmed the next Ascend 960DT chip is being pulled forward by three quarters, to Q1 2027. Huawei is targeting 1.6 million Ascend dies in 2026, more than double 2025 output.

The Performance Gap Still Matters

Huawei's newest Ascend 950PR delivers roughly 1.56 petaflops of FP4 compute versus about 20 petaflops for Nvidia's current Blackwell architecture, something like a 13x gap at the chip level. The Council on Foreign Relations estimates that even under aggressive assumptions, Huawei would account for only around 4% of Nvidia's aggregate AI compute output.

Huawei isn't catching up on raw capability. It's scaling fast enough to become the default option inside China regardless of that gap, which matters for how much of that market stays addressable for Nvidia at all.

Three Scenarios

Pulling the bull case, bear case and reverse-valuation math together, here's how the next two years could reasonably break down.

Bull, Base and Bear, With Explicit Assumptions

The bull scenario assumes hyperscaler capex keeps accelerating in line with what Meta, Microsoft and Amazon have already guided toward, and Nvidia holds its current share through the Blackwell and Vera Rubin transition. The market keeps paying a growth multiple because growth keeps showing up.

The base case assumes capex growth decelerates but doesn't reverse, more normalization than slowdown, with revenue growth settling into the mid-double-digits rather than 80%-plus. The multiple compresses toward where a maturing but still-dominant platform company tends to trade.

The bear scenario assumes one of the specific risks above actually materializes, a major customer visibly diversifying away, a sharp China policy reversal, or a broader capex pullback. None of these require Nvidia's technology to fail, only for the demand picture around it to weaken.

These outcomes mostly come down to how long the current growth rate holds and what the market keeps paying for it as that growth slows. The chart below lays the resulting ranges side by side.

At $226, NVDA sits above the base case's low end already, closer to "normalization" than bear-case territory. The market isn't pricing in a slowdown, it's betting the current pace mostly holds.

Trading the Debate

You don't need to resolve this argument before the next earnings report.

A stock CFD gives exposure to NVDA's price movement without ownership of the underlying shares, and leverage magnifies both gains and losses.

FAQ

Why is Nvidia's P/E so high?

It isn't, relative to its own history. NVDA's current trailing P/E near 28 sits well below its five-year average of around 60x.

What could make NVDA fall?

A slowdown in hyperscaler AI capex, a major customer diversifying away from Nvidia hardware, tighter export restrictions on China, or a broad reset in how much investors pay for AI-exposed growth stocks.

Is Nvidia overvalued compared to AMD?

By trailing P/E, no. AMD trades near 211x versus Nvidia's roughly 28x, though AMD's smaller base changes how that comparison reads.

What is Nvidia's PEG ratio and why does it matter?

Around 0.35, well under the 1.0 threshold often used as a rough fair-value marker, suggesting the price may be justified by expected growth rather than detached from it.

Is Huawei a real threat to Nvidia right now?

Not on raw performance, where Nvidia's lead is still roughly 13x in compute terms. It's a threat to Nvidia's addressable China market specifically.

Disclaimer: This content is for informational purposes only and should not be considered investment advice. Trading financial markets involves significant risk. Always conduct your own research before making any trading decisions.