NVIDIA Beat Everything and the Stock Barely Moved — That's the Buy Signal

By James Aspinwall | February 26, 2026

NVIDIA reported $68.1 billion in Q4 revenue (beat by $1.9B), $1.62 EPS (beat by $0.08), and guided Q1 to $78 billion (beat consensus by $5.2B — a 7% upside surprise). The stock moved 2% after hours, faded to 0.7% by the end of the earnings call, and closed the extended session roughly where it started.

A company that grew revenue 73% year-over-year, expanded gross margins, generated $35 billion in quarterly free cash flow, and guided to a number that annualizes to $340 billion — and the market shrugged.

This is either the market being efficient, or the market being wrong. Let’s figure out which.

Why the Stock Didn’t Move

1. Prediction Markets Had Already Priced the Beat

Prediction markets were 95% certain NVIDIA would beat earnings. Options traders priced in the smallest post-earnings swing in three years. When the beat is consensus, the beat isn’t news.

But there’s a difference between “will beat” and “by how much.” The Q1 guidance of $78B vs. $72.8B consensus is a massive 7.2% upside surprise on the forward number — the one that actually matters for stock prices. That gap should have moved the needle more than 2%.

2. AI Sector Rotation Headwinds

Bloomberg’s pre-earnings framing was telling: “Nvidia’s Stock Is So Stuck Even Blowout Earnings May Not Lift It.” The AI trade has been under pressure since the DeepSeek scare, with investors questioning whether hyperscaler capex can sustain its current trajectory. NVIDIA is being treated as a proxy for AI sentiment, not as a company with specific fundamentals.

When a stock becomes a macro trade, fundamentals get discounted.

3. The “What’s Next After Blackwell” Anxiety

The market is forward-looking past Q1 guidance. Vera Rubin ships H2 2026 — there’s a product transition gap where Blackwell is mature and Rubin isn’t yet generating revenue. Transitions create uncertainty, even when the roadmap is clear.

4. China Exclusion Narrative

NVIDIA explicitly excluded all Data Center compute revenue from China in Q1 guidance. That’s potentially $5-8 billion in annualized revenue written off. The market is treating this as permanent loss rather than conservative guidance.

Why the Market Is Wrong

The Math Doesn’t Support a Flat Price

Let’s run the numbers on what $78B in Q1 guidance actually implies:

Conservative FY27 estimate (assuming Q/Q growth decelerates to 10-12% through the year):

Quarter Revenue (est.) Rationale
Q1 FY27 $78.0B Guided
Q2 FY27 $85.0B Vera Rubin ramp begins
Q3 FY27 $92.0B Full Rubin production
Q4 FY27 $95.0B Seasonal + installed base
FY27 Total $350.0B +62% Y/Y

At 75% gross margins and current opex trends, that implies:

At $195-201/share, that’s a forward P/E of 25-26x on FY27 earnings.

For context: the S&P 500 trades at ~22x forward earnings. NVIDIA — growing revenue 62%, with 75% gross margins, 86% market share, and $97B in annual free cash flow — trades at roughly 3 P/E points above the market average.

That is not expensive. That is arguably cheap.

The PEG Ratio Is Screaming

Price/Earnings-to-Growth (PEG) ratio is one of the clearest value signals for growth stocks:

Metric Value
Forward P/E ~25-26x
Expected EPS growth ~60%
PEG ratio ~0.42

A PEG below 1.0 is considered undervalued. A PEG of 0.42 for a company of this quality is historically anomalous. For comparison:

NVIDIA’s PEG says the stock should be 2-3x more expensive relative to its growth rate, or the market expects growth to decelerate dramatically in FY28+.

Free Cash Flow Yield Is Attractive

At $4.7T market cap with $97B in FCF (and likely $130B+ in FY27), the FCF yield is:

That’s comparable to the 10-year Treasury yield. You’re getting a cash-flow machine growing 60%+ for roughly the same yield as a government bond. The risk-reward is asymmetric.

The Buyback Floor

NVIDIA has $58.5 billion remaining under its share repurchase authorization. They bought back $40 billion in FY26. At current prices, that represents ~2-3% of the float annually. This creates a structural demand floor — every dip gets bought by the company itself.

The Bear Case (And Why It’s Overstated)

“AI capex will peak”

This is the core bear thesis. If Microsoft, Meta, Google, and Amazon reduce AI infrastructure spending, NVIDIA’s revenue decelerates. But the evidence points the other direction:

The “peak capex” thesis has been wrong for four consecutive quarters.

“Custom ASICs will take share”

Google’s TPU and Amazon’s Trainium are real but captive — they only serve their own cloud. AMD is growing but from a small base. NVIDIA’s 86% market share has been stable for over a year despite all competitive efforts.

The $20 billion Groq deal eliminated the most credible inference-specific competitor. NVIDIA now owns both the throughput (GPU) and latency (LPU) sides of the inference market.

“The SBC change will hurt non-GAAP optics”

Including stock-based compensation in non-GAAP starting FY27 adds ~$1.9B/quarter to non-GAAP opex. This changes optics, not economics. The cash is still being generated. Analysts will adjust their models in one quarter and move on.

“China revenue is permanently lost”

The $0 China assumption in Q1 guidance is almost certainly conservative. NVIDIA has historically found ways to serve the China market with compliant products (H20, etc.). Even if China revenue is truly zero, the guidance already assumes this — any China revenue that materializes is pure upside.

The After-Hours Action Tells a Specific Story

The pattern — initial 3.5% pop, fade to 2.2%, settle near 0.7% — is characteristic of algorithmic de-risking, not fundamental repricing. Large institutional players who held through earnings take profits on the beat (their trade worked), and after-hours liquidity is too thin to absorb the selling.

This is mechanical, not informational. The after-hours market is 5-10% of normal volume. Prices set in thin liquidity are unreliable signals of fair value.

What matters is where the stock trades over the next 5-10 sessions as the full analyst community updates models with the $78B guidance number. The consensus price target will move from $257-269 higher. Institutional rebalancing will follow.

The Buy Case

Entry point: $195-201 (current range) 12-month target: $250-280 (based on 30-35x FY27 EPS of $7.80-8.00) Upside: 25-40% Catalyst timeline:

Risk/reward assessment: At a forward P/E of 25-26x with 60% earnings growth, the downside is limited (strong buyback floor, fortress balance sheet) and the upside is substantial. The stock would need to fall to ~$155 to reach a 20x forward P/E — that would require a fundamental thesis break, not just sentiment.

The muted earnings reaction isn’t the market telling you NVIDIA is fairly valued. It’s the market telling you it was already positioned for a beat and is now waiting for the next catalyst. The disconnect between the fundamental reality ($78B guidance, PEG of 0.42, FCF yield matching Treasuries) and the stock price creates an asymmetric opportunity.

When the best company in the fastest-growing market in technology trades at a PEG of 0.42, you buy it. The after-hours yawn is noise. The earnings report is signal.

Disclaimer: This analysis is for informational purposes only and does not constitute investment advice. The author may hold positions in securities discussed. Always do your own research and consult a financial advisor before making investment decisions.

Sources: