Most stocks don’t slip after beating estimates by this much. Nvidia did.

$81.6 billion in Q1 revenue. Up 85% year-over-year. EPS at $1.87 against a $1.76 consensus. A Q2 guide of $91 billion that assumes zero data center revenue from China. An $80 billion buyback. A dividend that went from $0.01 to $0.25 per quarter overnight. And shares traded down roughly 1% in the after-hours session. That reaction either tells you the market is exhausted by perfection, or it tells you the stock was already priced for all of it. Probably both.

So the real question on the table: does any of this make the stock cheap?

At first glance, no. Trailing P/E is sitting around 45.6x. The broader tech sector averages around 34x. That gap is real and it’s not nothing. If you handed someone those two numbers with no other context, they’d call it expensive and walk away. That’s the surface read, and it’s the one most casual observers stop at.

What matters is what happens when you move off trailing and look forward. The forward P/E on current estimates is approximately 26.9x. The semiconductor industry median is around 36x. That means Nvidia – after printing the biggest data center quarter in its history and guiding to $91 billion – is trading at roughly a 25% discount to its own industry peers on a forward earnings basis. At a $5.4 trillion market cap. That’s the sentence that deserves a second read.

And those forward estimates haven’t been revised yet. Wall Street updates models after prints, not before. A $91 billion guide with 75% gross margin is going to push consensus numbers higher over the next two to three weeks. Which means the 26.9x forward multiple is likely to compress further even if the stock doesn’t move.

There’s a PEG ratio sitting at 0.67 that not enough people are talking about. Below 1.0 historically implies a stock is undervalued relative to its growth rate. At 85% revenue growth, a number that low is unusual. It’s the kind of reading that shows up when earnings are outrunning price – not the other way around.

Slight tangent, but it matters here: NVDA’s 10-year average trailing P/E is approximately 66x. The stock is currently 31% below that historical average. Not marginally below – meaningfully below. And this is happening while the company carries $62.6 billion in cash, $11.4 billion in debt, $96.7 billion in trailing free cash flow, and a return on invested capital of 126%. The investors who paid 66x earnings for this company in prior cycles were buying a much smaller, less profitable business.

The bear case isn’t that the numbers are bad. It’s scale. A $5.4 trillion market cap creates compounding math problems. The stock hit an all-time high of $236.54 on May 14 – a week before earnings. Consensus price targets average around $275, implying roughly 23% upside from current levels, but that’s not a wide margin of safety on a stock this size. Jensen Huang has outlined a path to $1 trillion in combined Blackwell and Vera Rubin revenue across 2026 and 2027. If that trajectory holds, current valuations look conservative in hindsight. If demand softens anywhere in that chain – hyperscalers, enterprise, sovereign AI builds – 26x forward earnings on decelerating growth tells a different story fast.

Cheap in an absolute sense? No. Cheap relative to its own history, its growth rate, and its industry peers? Surprisingly, yes. The after-hours dip might be the most telling part of tonight – not because something is wrong, but because the market has run out of ways to be surprised. Whether that’s a ceiling or just a pause is the question nobody can answer cleanly yet.