Executive summary. In 2026 Jensen Huang organized NVIDIA around the AI factory, calling its buildout the largest infrastructure expansion in human history and saying it is accelerating. He put the six-chip Rubin platform into full production, said it cuts token costs to roughly a tenth of the prior platform, and argued the agentic AI inflection point has arrived.
The AI factory as the organizing idea
Huang's central 2026 message is that AI is now built in factories, and that building them is a historic undertaking. He called the buildout of AI factories the largest infrastructure expansion in human history and said it is accelerating at extraordinary speed. He tied that expansion to demand from a new workload, arguing the agentic AI inflection point has arrived and that enterprise adoption of agents is skyrocketing. At NVIDIA's earnings he repeated that agentic AI has arrived, doing productive work and scaling rapidly across companies and industries.
Rubin: betting the platform on cheaper tokens
The hardware expression of that thesis is Rubin, which Huang announced is now in full production as NVIDIA's first extreme-codesigned, six-chip platform and the successor to Blackwell. He said Rubin aims to push AI to the next frontier while slashing the cost of generating tokens to roughly one-tenth that of the previous platform, making large-scale AI far more economical to deploy. He framed that as a speed argument, saying the faster you train AI models, the faster you can get the next frontier out to the world. He also positioned NVIDIA's current generation as the cost-per-token leader, saying Grace Blackwell with NVLink delivers an order-of-magnitude lower cost per token and that Rubin will extend that leadership.
Full stack, from the data center to the road
Huang argued NVIDIA builds entire systems because delivering AI breakthroughs takes a full, optimized stack, saying its job is to create the whole stack so others can build applications for the world. At CES he applied that beyond the data center, launching Rubin alongside new open models for healthcare, robotics, and autonomy as he said AI is scaling into every domain and every device. He announced that the first passenger car with Alpamayo on NVIDIA's DRIVE platform would reach roads soon in the new Mercedes-Benz CLA, bringing AI-defined driving to the US this year. He highlighted momentum behind the level-4-ready DRIVE Hyperion platform, saying NVIDIA's vision is for every car and truck to eventually be autonomous.
Open and proprietary, and backing the frontier
On the open-source debate Huang rejected the binary, saying proprietary versus open is not a thing and that it is proprietary and open. He made the same point at GTC's session on open frontier models. He also positioned NVIDIA as a backer of the wider research frontier, announcing a significant investment in Thinking Machines Lab and calling AI the most powerful knowledge-discovery instrument in human history.
FAQ
What did Huang say about the Rubin platform?
He announced Rubin, NVIDIA's first extreme-codesigned six-chip platform and successor to Blackwell, is now in full production, and said it cuts token generation cost to about one-tenth of the previous platform.
How is Huang extending NVIDIA's stack into vehicles?
He announced the first car with Alpamayo on NVIDIA DRIVE arriving in the new Mercedes-Benz CLA and highlighted the level-4-ready DRIVE Hyperion platform.
How does Huang describe the AI-factory buildout?
He called it the largest infrastructure expansion in human history and said it is accelerating at extraordinary speed.
What is Huang's stance on open versus proprietary AI?
He said proprietary versus open is not a thing and that it is proprietary and open.
Sources
NVIDIA — blogs.nvidia.com (2026-01-05)
NVIDIA — blogs.nvidia.com (2026-03-25)
NVIDIA — investor.nvidia.com (2026-05-20)
NVIDIA — investor.nvidia.com (2026-02-25)
Thinking Machines Lab — thinkingmachines.ai (2026-03-10)
