NVIDIA Hits Hard at CES 2026
Jensen Huang delivered a dense keynote at CES 2026, unveiling a new generation of chips and architectures. From Vera CPU to Rubin and Alpamayo, NVIDIA clearly positions its ambitions: to dominate the era of agentic AI and autonomous vehicles.
Vera CPU: The Brain of Agentic AI
January 7, 2026 — NVIDIA launches Vera CPU, a processor designed specifically for AI reasoning.
| Feature | Value |
|---|---|
| Performance | 2x vs previous generation |
| Usage | AI reasoning, agents |
| Architecture | Control plane for agentic AI |
Introducing the NVIDIA Vera CPU. Announced at #CES2026, the #NVIDIAVera CPU is purpose-built for AI reasoning and delivers 2x the performance of the prior generation. Designed for accelerated systems, the Vera CPU is the control plane for agentic AI at scale. — @NVIDIADC on X
Rubin: Six Chips, A Giant Leap
January 6, 2026 — The Rubin architecture represents NVIDIA’s new frontier with six new chips optimized for training, inference, and advanced reasoning.
| Aspect | Description |
|---|---|
| Number of chips | 6 new ones |
| Design | Extreme codesign (compute, network, software) |
| Capabilities | Training, inference, reasoning at scale |
NVIDIA Rubin is designed to deliver unprecedented efficiency for both training and inference. Built with extreme codesign across compute, networking, and software, the platform delivers unprecedented efficiency for training, inference, and advanced reasoning at scale. — @NVIDIAAI on X
Runway x NVIDIA: Gen-4.5 on Vera Rubin
January 6, 2026 — Runway announces a major partnership with NVIDIA to run Gen-4.5 on the Vera Rubin architecture, even before its official release.
We’re thrilled to have partnered with @nvidia to bring Gen-4.5 to Vera Rubin - the first video generation model to run on NVIDIA’s most advanced accelerator, ahead of its release. Video generation and world models are ushering in a new era of AI. — @runwayml on X
This partnership positions Runway as a pioneer in video generation on next-generation hardware.
NVIDIA DRIVE AV: Mercedes CLA Premiere
January 6, 2026 — NVIDIA DRIVE AV debuts in the new Mercedes-Benz CLA with enhanced Level 2 autonomous driving capabilities.
| Feature | Detail |
|---|---|
| Level | Level 2 point-to-point |
| First vehicle | Mercedes-Benz CLA |
| Availability | US roads, late 2026 |
NVIDIA DRIVE AV software debuts in the all-new Mercedes-Benz CLA. At #CES2026, NVIDIA is bringing enhanced level 2 point-to-point driver assistance capabilities with expanded functionality to U.S. roads by the end of this year. — @NVIDIADRIVE on X
Nemotron Speech ASR: Ultra-Fast Transcription
January 6, 2026 — NVIDIA launches Nemotron Speech ASR, an open-source speech transcription model with record latency of 24ms.
| Feature | Value |
|---|---|
| Latency | 24ms |
| Type | Open-source |
| Usage | Real-time voice agents |
Open models are moving fast. Great to see Nemotron Speech ASR and the broader Nemotron stack pushing real-time voice AI forward—ultra low latency, full transparency, and production-ready paths for agents. This is exactly how open ecosystems scale impact. — @NVIDIAAI on X
Alpamayo: The AI That Reasons to Drive
January 5, 2026 — Jensen Huang announces Alpamayo, the first reasoning model designed specifically for autonomous vehicles. And it is open-source.
$NVDA CEO Jensen Huang just announced Alpamayo which he calls the world’s first thinking and reasoning model built for autonomous vehicles. By open sourcing the Alpamayo stack, Nvidia is pushing self driving forward as a category after years of work by thousands of engineers. — @StockSavvyShay on X
Open-sourcing Alpamayo marks a strategic turning point: NVIDIA bets on the ecosystem rather than exclusive ownership.
What This Means
Vera CPU redefines the role of the CPU in AI datacenters. It is no longer just a host for GPUs, but the “control plane” that orchestrates AI agents. The 2x performance targets reasoning and multi-agent orchestration workloads directly.
Rubin with its 6 chips represents NVIDIA’s “extreme codesign” approach: hardware, network, and software optimized together. This is the answer to scaling limitations observed in recent years.
The partnerships (Runway, Mercedes) demonstrate that NVIDIA is not just selling hardware—the company is building complete vertical ecosystems.
Open-source (Nemotron Speech, Alpamayo) marks a strategic shift. NVIDIA understands that domination comes through massive adoption, not closure. By opening its models, it creates de facto standards.