Every major analyst on Wall Street is telling you to buy $NVDA before February 25. Thirty-seven Buy ratings. Zero Sells. An average price target of $260 — 42% above where the stock sits today. That kind of consensus should make you nervous.
Nvidia reports Q4 FY2026 earnings next Tuesday after the bell. The numbers Wall Street expects are staggering: $65.6 billion in revenue (up 67% year-over-year) and earnings per share of $1.52 (up 71%). For context, that quarterly revenue figure is larger than Intel's entire annual revenue. In a single quarter.
And yet — the stock is basically flat in 2026. Down about 2% while the S&P has pushed higher. Something doesn't add up.
The Setup
Last quarter was supposed to be the blowout that sent $NVDA to new highs. It wasn't. Revenue hit $57 billion, beating estimates by $2 billion. EPS came in at $1.30, ahead of the $1.25 consensus. The stock dropped 3% the next day.
The problem isn't the numbers. The numbers are absurd. The problem is expectations have gotten so high that beating them doesn't move the needle anymore.
| Metric | Q3 FY2026 (Actual) | Q4 FY2026 (Expected) | YoY Growth |
|---|---|---|---|
| Revenue | $57.0B | $65.6B | +67% |
| EPS | $1.30 | $1.52 | +71% |
| Data Center Revenue | $51.2B | $59.9B | +66% |
| Gross Margin | 73.6% | 75.0% | Expanding |
| Gaming Revenue | $4.3B | $4.1B | Flat |
Data center is the whole story. It's 90% of revenue now, up from about 80% a year ago. Gaming, the business that built Nvidia, is an afterthought.
The Bull Case
It's simple: everyone is spending and nobody is slowing down.
Alphabet just announced $175-185 billion in 2026 capex — almost all of it going to AI infrastructure. Amazon committed $200 billion. Meta, Microsoft, same story. These hyperscalers are Nvidia's top customers, and they're locked into a GPU arms race with no end in sight.
The new Rubin chips give Nvidia pricing power through 2027. CUDA keeps customers locked into the ecosystem. And at 25x forward earnings, $NVDA is actually cheaper than Alphabet (28x) and Broadcom (34x) despite growing three times faster.
CHART_PLACEHOLDER_0
Wolfe Research's Chris Caso raised his FY2028 EPS estimate to $11.50 — that's $1.50 above consensus. If he's right, $NVDA at $183 is a steal.
The Bear Case
Two words: inventory buildup.
Seeking Alpha flagged elevated inventory levels last week. When a company growing 67% per year is building inventory faster than revenue, it's worth asking why. Either demand is about to accelerate even further (Nvidia's argument) or customers are starting to absorb what they've already ordered before placing new ones.
Then there's DeepSeek. The Chinese AI lab proved you can train competitive models with far fewer GPUs. That doesn't kill Nvidia's business — inference demand is exploding regardless — but it chips away at the narrative that you need unlimited compute to compete in AI.
And the stock's reaction to the last three earnings tells you something. Beat, beat, beat — and the stock went nowhere. At some point, "good" has to become "great enough to justify the most crowded trade on the planet."
What to Watch on Feb 25
Forget the headline numbers. They'll beat. They always beat. Here's what actually matters:
Q1 FY2027 guidance. UBS expects Nvidia to guide to $76 billion. If it's $74 billion or below, the stock sells off regardless of Q4 results. This is a guidance stock now.
Gross margins. Q3 was 73.6%. Wall Street wants 75%+. If Blackwell ramp costs compress margins, that's a problem.
Customer concentration. How much revenue came from the top 5 hyperscalers vs. enterprise and sovereign AI? Broadening demand is bullish. Concentration is a risk.
China revenue. AMD's recent results suggested a few billion in incremental China GPU sales. If Nvidia saw the same, it's upside the market hasn't priced in.
CHART_PLACEHOLDER_1
Bottom Line
Nvidia will almost certainly beat Q4 estimates. That's not the question. The question is whether guidance and gross margins are strong enough to break the stock out of a range it's been stuck in for six months.
At 25x forward earnings with 60%+ growth, the valuation is defensible. But the trade is crowded, expectations are sky-high, and the last three earnings produced sell-the-news reactions. If you're buying ahead of the print, you're betting this time is different. Sometimes it is. But the bar has never been higher.
NVDA