Oracle CEO announces $100+ billion investment in 2000+ datacenters, dedicating 40% to Nvidia GPU infrastructure.

Oracle is pursuing an aggressive expansion into the data center market, with CEO Larry Ellison announcing the company will spend $100 billion over the next 4 years, with NVIDIA set to receive 40% of that investment. Currently, Oracle operates 162 cloud data centers in operation and under construction across the world. Ellison stated that "the largest of these datacenters is 800 megawatts and will contain acres of NVIDIA GPU Clusters for training large scale AI models." Looking forward, Oracle could operate up to 2,000 data centers in the future, a significant expansion from its current footprint.

The scale of Oracle's ambitions is reflected in the infrastructure it is designing to support AI workloads. The company is developing a data center with capacity exceeding a gigawatt, powered by three small modular nuclear reactors. Ellison explained the unprecedented scale of these efforts: "This is how crazy it's getting. This is what's going on." The nuclear reactors, for which building permits have already been obtained, underscore the massive power demands required to run frontier AI systems.

Ellison emphasized that the barrier to entry for competing in frontier AI development remains extraordinarily high. He stated: "these frontier models are going to -- the entry price for a real frontier model from someone who wants to compete in that area is about $100 billion. Let me repeat, around $100 billion. That's over the next 4, 5 years for anyone who wants to play in that game." Ellison added that this cost structure means "there are not going to be a lot of those" frontier model developers.

Beyond foundational models, Ellison highlighted the emergence of highly specialized AI applications. He noted work "using computers to look at, biopsies of slides or CAT scans to discover cancer" as well as "blood tests for discovery and cancer," which are trained on domain-specific data such as "literally millions of biopsy slides." These specialized models differ from foundational platforms like ChatGPT and Gemini in their narrow focus and training data requirements.

Ellison concluded that the AI infrastructure race shows no signs of slowing. He stated: "I think this is an ongoing battle for technical supremacy that will be fought by a handful of companies and maybe one nation state over the next 5 years at least, but probably more like 10. So this business is just growing larger and larger and larger. There's no slowdown or shift coming." The investment demands and technical requirements will continue to intensify across both frontier and specialized model development.