Oracle’s AI Bet is Either Genius or a Trainwreck — And We’ll Find Out Soon

If you want to know whether the AI bubble is actually popping, don’t look at Nvidia or OpenAI. Look at Oracle.

Yes, the boring database company that your grandpa’s company still uses for payroll. Larry Ellison’s crew has done a full scorched-earth pivot to AI, but not in the way you’d expect. They’re not building foundation models like OpenAI or Anthropic. They’re not quite a neocloud like CoreWeave either, though they’ve jumped into the same bare-metal server business.

What Oracle is doing is something weirder: a legacy SaaS company making an audacious, almost reckless bet on a very specific future for AI infrastructure. And because Oracle is publicly traded, their earnings calls are basically a real-time stress test for the entire AI hype cycle.

The bet, in simple terms

Oracle is renting out massive clusters of GPUs and networking gear to AI companies. Not just renting — they’re building purpose-built data centers designed for the insane power and cooling demands of training large models. They’re competing head-to-head with cloud giants like AWS, Azure, and Google Cloud, but on a narrower playing field. Their pitch? We’ll give you raw compute capacity faster and cheaper than the hyperscalers, without all the lock-in nonsense.

It’s a high-risk, high-reward play. The infrastructure buildout costs billions upfront. If AI demand keeps growing, Oracle wins big. If the market cools off or the big model builders start optimizing their own hardware, Oracle is left holding a mountain of expensive GPUs that nobody wants.

Why Oracle is the bellwether

The Verge’s piece calls Oracle “the only publicly traded company that will tell you” if the AI bubble is bursting. I think that’s overselling it a bit — plenty of other companies are exposed — but there’s truth to the idea. Oracle’s AI business is pure, undiluted exposure to the AI infrastructure buildout. They don’t have a consumer product or a search engine or a social network to fall back on. If AI demand stumbles, Oracle’s stock gets crushed. There’s no hedge.

Compare that to Microsoft, which has Azure AI but also Office, Windows, LinkedIn, and gaming. Or Google, which has search ads, YouTube, and Android. Oracle’s traditional database business is in slow decline. This AI pivot is their whole future.

The Larry Ellison factor

Ellison is 81 years old and still running the show. The guy has always been a gambler — he famously bet the company on database software in the 70s when everyone else was building mainframes. He’s also known for being… let’s say, creatively optimistic in his public statements. His earnings call pronouncements about AI demand should probably be taken with a grain of salt.

But here’s the thing: Ellison has been right before. A lot. And he’s putting his own money where his mouth is. Oracle’s AI infrastructure spending is real, the contracts with companies like Cohere and xAI are real, and the revenue is starting to show up in their financials.

The risk nobody’s talking about

Most analysts focus on whether AI model training demand will hold up. I think the bigger risk is something else: the race to build custom AI chips. If OpenAI, Google, Amazon, and Meta all succeed in making their own silicon that’s dramatically more efficient than Nvidia’s GPUs, the entire market for rented Nvidia clusters could shrink fast. Oracle is essentially betting that Nvidia’s dominance will last long enough for them to recoup their investment.

That’s a bet on both AI demand and Nvidia’s continued moat. Two bets stacked on top of each other.

What I’m watching

Oracle’s next earnings report is going to be a big one. I’ll be looking at their AI infrastructure revenue growth, their forward guidance on capital expenditures, and any hints about customer concentration. If one or two big customers represent most of their AI revenue, that’s a red flag. If demand is broad and growing, that’s a green light for the whole sector.

Either way, it’s going to be interesting. Oracle has always been the boring enterprise company that somehow survives every tech wave. This time they’re not just surfing the wave — they’re betting the whole company on it.