After four straight quarters of beating expectations and watching the stock sell off anyway, Nvidia finally broke the pattern. Shares jumped roughly 7% on August 27, adding about $359 billion in market value in a single session. The catalyst: quarterly revenue of $96.2 billion — up 106% year over year — paired with guidance so aggressive it made Wall Street’s prior models look outdated before the ink dried.
The Numbers Don’t Whisper
The gap between what analysts expected and what Nvidia delivered is wide enough to drive a data center through.
Data center revenue hit approximately $89 billion, up 117% year over year. The ACIE segment — AI Clouds, industrial, and enterprise customers — generated roughly $40.3 billion, up about 138%. That last number matters: AI spending is no longer a hyperscaler-only story.
- Q2 revenue of $96.2B beat analyst expectations in the low-$90B range
- Data center: ~$89B, up ~117% YoY; ACIE up ~138% YoY
- Fiscal 2028 guidance: ~70% revenue growth vs. analyst consensus of ~44–45%
- Implied FY2028 revenue: ~$690–700B versus the ~$570B prior consensus
CFO Colette Kress issued a rare full-year outlook — and the number was 70% growth, not the 44–45% analysts expected. Nvidia frames this as supply-constrained, meaning TSMC capacity and high-bandwidth memory shortages are the ceiling, not weak demand. One portfolio manager told CNBC there’s “plenty, plenty of upside” in the shares given the trajectory.
Jensen Huang called this AI’s “inflection point” — multiple frontier labs building simultaneously, a thriving open-source ecosystem, physical AI embedded in robots and edge devices gaining momentum. Whether you find that inspiring or exhausting probably depends on how many AI feature announcements you’ve absorbed this year.
Nvidia’s Reported $13 Billion Software Land Grab
According to reports, Nvidia is moving to control not just the chip, but the platform developers build on.
According to reports from The Information and Business Insider, Nvidia has reportedly agreed — or is in advanced talks — to acquire Hugging Face for approximately $12.9–13 billion. Nvidia has not formally confirmed the deal.
Hugging Face is the dominant open-source platform for hosting and deploying AI models. The platform generates roughly $150 million in annualized revenue — meaning Nvidia is reportedly paying about 86 times revenue. That’s ecosystem capture, not a financial return play.
If the deal closes, open-source AI development gets tighter integration with Nvidia’s hardware stack. Whether that means better performance or gradual lock-in depends entirely on how much you trust a company that already controls the dominant AI chip platform. A deal of this size would also likely draw antitrust scrutiny in the U.S. and EU, though no specific regulatory action has been reported yet.
One technology analyst told CNBC there’s “no sign of a slowdown until at least 2028.” The bull case is compelling. So was every other supercycle story right before it wasn’t. Chip stocks shed roughly $1 trillion in July. Hyperscalers and OpenAI are quietly designing custom AI chips to reduce Nvidia dependence. The monoculture risk is real.
The supply constraints — TSMC capacity, high-bandwidth memory shortages — remain the actual ceiling. Demand isn’t Nvidia’s problem. Shipping enough hardware to hit 70% growth is the only number worth watching now.
