Nvidia CEO Jensen Huang gives keynote address at CES, unveils automated vehicle AI

NVIDIA at CES: Vera Rubin, Autonomous Driving, and Future Innovation

Key Concepts:

• Vera Rubin: NVIDIA’s next-generation chip architecture, comprising six chips, designed for significantly improved price-performance in AI inference.
• Blackwell: The current generation of NVIDIA chips, serving as a benchmark for Vera Rubin’s performance improvements.
• AI Inference: The process of using a trained AI model to make predictions or decisions.
• Autonomous Driving: Development of self-driving vehicle technology, with NVIDIA focusing on reasoning capabilities beyond basic traffic rule adherence.
• Price-Performance Ratio: A measure of the computational power delivered per dollar spent on hardware.

NVIDIA’s CES Keynote & Vera Rubin Announcement

The core of NVIDIA’s presentation at CES revolved around “Vera Rubin,” their upcoming chip architecture. All six chips comprising Vera Rubin are now in full production, with partner systems expected to arrive in the second half of the year. A key statistic highlighted was a “ten times lower inference token cost versus Blackwell,” indicating a substantial improvement in price-performance. NVIDIA is positioning Vera Rubin as a premium offering, arguing that the increased efficiency justifies the potential higher cost. The company emphasized backward compatibility, making it relatively easy for users to upgrade from previous generations. Runway, an AI video company partnered with NVIDIA, reportedly transitioned to Vera Rubin in a single day and immediately experienced the performance benefits. This demonstrates NVIDIA’s commitment to continuous innovation and maintaining a competitive edge, not just in performance but also in efficiency.

Autonomous Driving Advancements

NVIDIA also announced advancements in autonomous driving software, building upon their existing work in the field. The new system incorporates enhanced “reasoning capabilities,” allowing vehicles to navigate situations beyond pre-programmed rules. An example cited, though not directly attributed by NVIDIA, was the challenge Waymo cars faced during the San Francisco blackout when traffic lights were non-functional. The implication is that vehicles equipped with NVIDIA’s new technology would be able to make independent decisions in such scenarios. NVIDIA representatives were careful to avoid direct comparisons to Waymo, but the example illustrates the potential for improved autonomous navigation in complex, unpredictable environments.

Audience Reaction & Analyst Questions

Following the keynote, Jensen Huang engaged in a Q&A session with analysts. The reporter anticipates gaining insights from these questions, as well as conducting his own interview with Huang the following morning on Squawk Box. The reporter noted the Q&A was taking place approximately 50 feet from his location, suggesting a high level of interest and engagement from the analyst community.

Data & Statistics

• 10x: The reported reduction in inference token cost with Vera Rubin compared to Blackwell.
• Second Half of the Year: Expected arrival timeframe for partner systems utilizing Vera Rubin.

Logical Connections

The presentation demonstrates a clear strategy from NVIDIA: to maintain leadership in the AI hardware space through continuous innovation. The Vera Rubin announcement directly addresses the need for improved price-performance, while the advancements in autonomous driving showcase the application of this technology in a critical and growing market. The emphasis on backward compatibility is a strategic move to encourage adoption and minimize disruption for existing customers.

Notable Quotes

While no direct quotes from Jensen Huang are included in this segment, the reporter relays the sentiment from the CEO of Runway, who “was singing its praises” regarding the benefits of switching to Vera Rubin.

Technical Terms

• Inference Token Cost: A metric related to the computational resources required to process data using an AI model. Lower cost indicates greater efficiency.
• Backward Compatibility: The ability of newer hardware or software to work with older systems.

Conclusion

NVIDIA’s CES presentation underscored its commitment to pushing the boundaries of AI technology. The Vera Rubin announcement represents a significant leap forward in price-performance, while the advancements in autonomous driving highlight the practical applications of NVIDIA’s innovations. The company is positioning itself not just as a hardware provider, but as a key enabler of the future of AI and autonomous systems. The upcoming interview with Jensen Huang promises to provide further details and insights into NVIDIA’s strategic direction.