Nvidia closed at $208.27 on Friday, up 4.3%, taking out the October 29 peak and crossing $5 trillion in market value. The skeptical read is that the move is fragile — propped up by Intel’s earnings, vulnerable to next week’s hyperscaler reports. The skeptical read has it backward. The fundamentals were in place months ago. The price has finally caught up.
The numbers Nvidia has already put on the board do most of the work. Q4 fiscal 2026 revenue of $68.1 billion, up 73% year over year. Data center revenue of $62.3 billion, up 75%. Adjusted EPS of $1.62, the fifth consecutive quarterly beat. Free cash flow of $96.7 billion. Q1 FY2027 guidance of $78 billion, implying roughly 77% growth — and that figure excludes China data center compute revenue entirely. At Friday’s close, the stock trades at a forward P/E of around 24x. A company compounding at 73% on that scale of cash generation does not normally clear that multiple. The compression was the anomaly. The breakout is the correction.
Intel’s print did not lift Nvidia. It confirmed what Nvidia’s order book already implied. Intel’s data center chip revenue grew 22% year over year — the kind of broadening that signals AI infrastructure spending is not a one-vendor distortion but a multi-vendor cycle, with Nvidia at the center of it. That tape pulled the Philadelphia Semiconductor Index to an eighteenth straight session of gains, with the index up 47%. The sector is trending because the demand is real, not because one earnings report papered over weakness.
Next week’s hyperscaler reports are a catalyst, not a risk. Microsoft, Amazon, Alphabet, and Meta are collectively guiding toward roughly $700 billion of 2026 capital spending. The pattern at every prior print in this cycle has been the same: capex guidance gets revised up, not down. Meta’s outlook has climbed every quarter. Microsoft has repeatedly signaled Azure AI capacity remains supply-constrained. Alphabet’s own silicon program is real but does not displace Nvidia within the time horizon the stock is currently pricing. The probability skew on these reports favors extension, not reversal.
The forward order book is the strongest single piece of evidence. At GTC in March, Huang disclosed at least $1 trillion in committed Blackwell and Vera Rubin orders through end-2027 — double the $500 billion figure cited at the same event a year earlier. Vera Rubin, shipping in the second half of 2026, is engineered for roughly 10x the performance per watt of Grace Blackwell, which is the metric hyperscalers care about most as power becomes the binding constraint on data center expansion. The H200 production line for China has reopened under U.S. export licenses, restoring a revenue stream the market had already marked to zero. KeyBanc maintains an Overweight rating with a $275 price target and treats the well-publicized HBM4 timing slippage on Rubin as a contained issue against a 60,000-plus NV72 rack shipment year on existing Blackwell. The buffer is structural.
What the six-month consolidation reflected was not a problem with the business. It was a market waiting for permission to re-rate. Order book, free cash flow, growth rate, customer capex — every fundamental input moved up while the price went sideways. Friday delivered the tape strong enough to force the gap closed.
Nvidia didn’t break out. It caught up.