Hey there, bargain hunter. Nvidia dropped its most jaw-dropping quarter in history last week — and the stock dipped anyway. If that sentence doesn’t make you stop and think, nothing will.
Let’s run the numbers first, because they genuinely deserve a moment of silence.
The scoreboard: Q1 FY2027 revenue hit $81.6 billion, up 85% year-over-year and 20% sequentially. Data Center revenue — the engine of everything — came in at $75.2 billion, up 92% from a year ago. Net income was $58.3 billion. Non-GAAP EPS landed at $1.87, blowing past the $1.77 Wall Street consensus. Q2 guidance came in at $91 billion — above the Street’s $86 billion target. And Nvidia announced an $80 billion share repurchase alongside a 25x dividend hike.
By any rational measure, this is a company operating at a scale that almost no business in history has ever achieved. Jensen Huang put it plainly on the earnings call: demand has gone parabolic. The reason, he said, is simple — agentic AI has arrived.
So why did the stock slip roughly 1.5% in after-hours?
Here’s the part people skip. Wall Street had, for the first time in this cycle, actually pulled its estimates above Nvidia’s own guidance heading into the print. The bar had quietly been raised to an almost impossible level. A $91 billion Q2 guide is extraordinary — but analysts had whispered $86B as the floor. When you price in perfection, a merely spectacular result reads as a disappointment.
Slight tangent, but it matters: Nvidia’s customer base is quietly diversifying in a way that rarely gets attention. Hyperscalers — Microsoft, Google, Amazon, Meta — used to account for well over half of all data center revenue. They still matter enormously. But a new bucket called ACIE (AI Clouds, Industrial, Enterprise) now represents nearly equal revenue. Sovereign AI deployments tripled year-over-year to over $30 billion in fiscal 2026. That diversification reduces the single-hyperscaler pullback risk that bears have leaned on for two years.
The networking business also deserves a mention. Data center networking revenue — InfiniBand, Spectrum-X Ethernet, NVLink — came in at a record $14.8 billion, up 199% year-over-year. That’s not a chip company. That’s a platform.
Is It Cheap?
At roughly 30.5x calendar 2026 estimated earnings, NVDA is not screaming value. It’s pricing in several years of continued dominance. The stock has rallied over 150% from its April 2025 low. At a $5.4 trillion market cap, it is now estimated to account for roughly 50% of S&P 500 performance in certain periods — which is its own kind of concentration risk.
But here’s where it gets interesting for the bargain hunter. The $91 billion Q2 guide includes zero China revenue. Meaning the business grew to this scale with a major market effectively shut off. If export restrictions ever ease — even partially — that’s pure upside nobody is modeling right now.
Bull case: Vera Rubin architecture ramps in late 2026, the new CPU business opens a $200 billion TAM, and agentic AI drives inference demand that keeps the data center flywheel spinning well into 2028.
Bear case: At 30x forward earnings, any deceleration reads as disaster. Competition from custom ASICs, in-house hyperscaler chips, and AMD’s next-gen ramp could erode margins faster than the current setup implies.
Base case: Nvidia stays the dominant AI infrastructure platform through at least 2027. But the easy money was made at $86. From here, the stock moves on whether the $91 billion guide proves conservative — or just right.
The dip after a monster beat is a classic setup. Whether it holds is a different question entirely.
