In a historic announcement that could redefine the trajectory of artificial intelligence, NVIDIA and OpenAI have unveiled what they are calling the largest AI infrastructure deployment in history. This partnership is set to scale OpenAI’s compute capabilities to levels never seen before, fueling breakthroughs that will shape virtually every industry.
A Strategic Partnership to Scale AI Globally
During a discussion with CNBC’s Jon Fortt, Jensen Huang (Founder & CEO of NVIDIA), Sam Altman (CEO of OpenAI), and Greg Brockman (President of OpenAI) described their shared vision for an AI-powered future.
“This is the biggest AI infrastructure project in history,” Huang emphasized. “This partnership is about building an AI infrastructure that enables AI to go from the labs into the world.”
Through this collaboration, OpenAI will deploy at least 10 gigawatts of NVIDIA systems, powered by the NVIDIA Vera Rubin platform. To support the scale-up, NVIDIA will invest up to $100 billion into OpenAI, linked to each gigawatt of deployment.
Building Million-GPU AI Factories
The agreement will create AI factories with millions of GPUs, designed to meet the unprecedented demands of training and inference for next-generation AI models.
“This is the fuel that we need to drive improvement, better models, revenue — everything,” Altman said. “Without this infrastructure, we’d have to choose between transformative use cases like cancer research or global education. No one wants to make that choice.”
For perspective: OpenAI’s ChatGPT, launched in 2022, became the fastest app to reach 100 million users. Today, it serves over 700 million weekly active users. Supporting this scale — and the new frontiers of agentic AI, reasoning models, multimodality, and longer context windows — requires compute at a level only this partnership can provide.
From 2016 to 2026: A Billion-Fold Leap
This isn’t the first time the two companies have joined forces. In 2016, Huang hand-delivered the first NVIDIA DGX system to OpenAI’s San Francisco headquarters.
Fast forward a decade: the first gigawatt-scale Vera Rubin GPUs will generate their tokens in the second half of 2026.
“This is a billion times more computational power than that initial server,” Brockman said. “It unlocks the ability to create new breakthroughs, new models, and empower every individual and business.”
The Beginning of a Massive AI Buildout
Huang underlined that this project is just the beginning:
“We’re literally going to connect intelligence to every application, every use case, every device — and we’re just at the start. This is the first 10 gigawatts, I assure you of that.”
With trillion-parameter models, agentic AI, and reasoning capabilities becoming mainstream, the partnership signals not just a hardware expansion, but the dawn of AI as an infrastructure for society, akin to electricity or the internet.
Why This Matters
- For Industry: Businesses worldwide will soon have access to AI models more powerful, scalable, and affordable than ever before.
- For Science & Medicine: With near-unlimited compute, AI can simultaneously drive breakthroughs in healthcare, climate science, and beyond.
- For Users: Everyday people will experience AI assistants that are smarter, faster, and more context-aware.
The NVIDIA–OpenAI alliance marks a tipping point — moving AI from niche labs to global-scale deployment that touches every corner of life.