CES: Nvidia’s Rubin Cuts AI Inference Costs 10×
NVIDIA Rubin GPUs slash AI inference costs 10x, debut at CES 2026.
“AI Disruption” Publication 8500 Subscriptions 20% Discount Offer Link.
“Every 10 to 15 years, the computing industry undergoes a revolution, each time giving birth to new forms of platforms. Now, two transformations are happening simultaneously: applications will be built on AI, and the way you build software will also change.”
Just this morning, at the CES 2026 exhibition in Las Vegas, NVIDIA founder Jensen Huang appeared wearing his signature leather jacket!
The first slide Huang displayed read: “The development of artificial intelligence has surpassed large language models.”
With advances in large language model technology, future physical world AI will be able to understand the structure of the real world, complete tasks independently, and learn over time. He stated that “anywhere in the universe where information exists, where structure exists” can be used to train artificial intelligence.
Huang shared how next-generation accelerated computing and artificial intelligence will transform every industry, and introduced NVIDIA’s latest developments in chips, artificial intelligence models, open source, and other areas, mainly including:
• Next-generation Rubin platform;
• New Vision-Language-Action model (VLA) — Alpamayo 1;
• New open models, frameworks, and AI infrastructure for physical AI.
This includes not only next-generation GPUs but also industry-leading open-source AI models. It’s clear that by 2026, NVIDIA is preparing to lead technological development in a full-stack manner.
