According to Inc, Nvidia reported $57 billion in revenue yesterday, blowing past Wall Street’s $55 billion expectation. The company’s data center division specifically delivered $51 billion, beating the $49.31 billion analyst prediction. CEO Jensen Huang rejected talk of an AI bubble during the earnings call, arguing we’re seeing a longer transformation instead. Nvidia forecast $65 billion in revenue for the fourth quarter, while analysts had only expected $61.98 billion. Wedbush analyst Daniel Ives called this a “1996 moment” rather than a 1999-style bubble, suggesting we’re entering “Year 3 of a 10-year build out.” Another analyst, Thomas Monteiro from Investing.com, said the data shows AI is “nowhere near its peak.”
The 1996 Moment
Here’s the thing about Ives’ “1996 moment” comparison – it’s actually pretty smart. 1996 was when the internet was real but hadn’t gone mainstream crazy yet. Companies were building infrastructure, not just hyping concepts. That’s exactly where we are with AI right now. Nvidia isn’t selling dreams – they’re moving actual hardware at staggering scale. The data center numbers don’t lie. When you’re beating expectations by billions quarter after quarter, something fundamental is happening.
Beyond the Hype
Huang made an interesting point about people eventually looking deeper at the technology rather than just focusing on CapEx. He’s basically saying we’re still in the “wow, this is magic” phase. But soon, businesses will need to actually make this stuff work efficiently. That’s where the real transformation happens. And honestly, when you look at companies that need reliable industrial computing power – like those using industrial panel PCs for manufacturing automation – the demand for robust AI infrastructure becomes obvious. This isn’t just about chatbots writing poems.
Scaling Isn’t Optional
Monteiro nailed it when he said datacenter scaling “is not optional, but rather the central need for every tech business in the world.” Think about that for a second. We’re not talking about nice-to-have upgrades anymore. AI infrastructure is becoming as essential as electricity or internet connectivity. When Nvidia’s CEO talks about transitions to accelerated computing and agentic AI, he’s describing a fundamental shift in how computing works. The companies that get left behind here won’t just be slower – they’ll be obsolete.
What Comes Next
So where does this go from here? If we’re really in year three of a ten-year buildout, we haven’t seen anything yet. The applications are just starting to emerge beyond the obvious use cases. Physical AI – robots, autonomous systems, smart manufacturing – that’s where things get really interesting. The demand for computing power in industrial settings alone could dwarf what we’ve seen so far. The bubble talk feels increasingly disconnected from the actual numbers. When a company consistently beats expectations by billions, maybe the problem isn’t the company – it’s the expectations.
