NVIDIA CES 2026: Rubin, DLSS 4.5 and the future of Physical AI

Jensen Huang opened CES 2026 with a bold statement: “Computing has been fundamentally reshaped by accelerated computing and AI”. The announcements that followed backed up that claim.

Rubin: The 6-Chip AI Platform

NVIDIA unveiled Rubin, its first extreme-codesigned 6-chip AI platform. This architecture represents a generational leap in how AI systems are designed:

Feature	Detail
Architecture	6 codesigned chips
Status	Full production
Focus	Enterprise-scale AI

The philosophy behind Rubin is clear: instead of just scaling chip size, NVIDIA is optimizing how multiple chips work together natively.

DLSS 4.5: Next-Gen AI Gaming

For game developers and graphics application creators, DLSS 4.5 brings significant improvements:

Dynamic Multi Frame Generation

New 6X Multi Frame Generation mode
Second-generation transformer model for Super Resolution
Over 250 games and apps now support DLSS 4

DLSS Evolution:
DLSS 1.0 → Basic Super Resolution
DLSS 2.0 → Improved model
DLSS 3.0 → Frame Generation
DLSS 4.0 → Multi Frame Generation
DLSS 4.5 → Dynamic MFG + 6X mode

Implications for developers

If you’re developing graphics-intensive applications:

Simpler integration: The DLSS SDK continues to improve ease of use
Multiplied performance: The 6X mode can transform the experience on mid-range hardware
Massive adoption: With 250+ titles supporting the technology, it’s a de facto standard

Robotics: NVIDIA wants to be the default platform

The most significant announcement for the future of software development was the complete ecosystem for Physical AI:

What NVIDIA presented

Component	Description
Foundation Models	Open models for robotic reasoning
Alpamayo	Model family for autonomous vehicles
Simulation Tools	Tools for training robots virtually
Edge Hardware	Optimized hardware for on-device execution

Why it matters

NVIDIA aims to become the default platform for generalist robotics

This means developers who want to work on:

Humanoid robots
Autonomous vehicles
Industrial automation
Intelligent drones

…will have a complete NVIDIA stack to do so.

Competition comparison

At CES 2026, there was a historic moment: Jensen Huang, Lisa Su (AMD) and Lip-Bu Tan (Intel) shared the stage at Lenovo’s event.

Company	Physical AI focus
NVIDIA	Complete platform (models + hardware + simulation)
AMD	GENE.01 humanoid robot (with Generative Bionics)
Intel	RoBee humanoid robot (with Oversonic Robotics)

NVIDIA is the only one offering a complete stack for developers.

Implications for businesses

Immediate opportunities

Graphics application development: DLSS 4.5 reduces required hardware costs
Robotics exploration: Open foundation models allow experimentation
Autonomous vehicles: Alpamayo offers a standardized entry point

Considerations

Vendor lock-in: The NVIDIA ecosystem is powerful but creates dependency
Hardware investment: Rubin will be expensive initially
Learning curve: New development paradigms

Next steps

If your company is evaluating Physical AI or game/3D application development:

Evaluate DLSS 4.5 if you already work with real-time graphics
Explore the robotics foundation models if you have automation use cases
Monitor Rubin availability to plan future investments

Want to explore how these technologies can benefit your company? Contact us for an AI opportunities assessment.