Oracle has announced that it will be spending "billions" of dollars on NVIDIA's most advanced AI chips to power its cloud services, with Oracle chairman Larry Ellison announcing the news at a recent Ampere event.

The company will "cash on" the AI hype through NVIDIA's super-powerful AI GPUs while using CPUs that it will secure from both Ampere and AMD. Oracle says it will spend "three times more" money on Ampere and AMD CPUs than it will on NVIDIA GPUs, guaranteeing the company will be dominant in the cloud industry with AI.

But what about Intel? 

Oracle has said that the Intel x86 architecture is "old" and has reached its limits and that it is more attracted to building out a relationship with AMD and Ampere for its CPU needs while using NVIDIA for its GPU and AI needs. Intel has been a long-term partner of Oracle, rewinding the clock back to 2008 when the company first released its Exadata machine.

Oracle confirmed back in October 2022 that it would be using many NVIDIA high-end A100 and H100 computing GPUs, which formed the relationship between NVIDIA and Oracle. But now, it seems Oracle is taking a giant leap into NVIDIA's warm AI-powered GPU arms.

As it stands, 16,000+ NVIDIA AI GPUs are installed into the very latest cloud data centres at Oracle, but those numbers will only grow massively as the company pumps billions of dollars into new NVIDIA AI GPUs in the near future.