NVIDIA CEO to Introduce Revolutionary Chips at GTC 2023, Hinting at Rubin or Next-Gen Feynman AI Series

NVIDIA’s CEO recently provided exciting insights into what we can anticipate at the upcoming GTC 2026 event, suggesting that attendees may witness the introduction of unprecedented chips.

Anticipated Innovations: Next-Gen Feynman Chips and Groq’s LPU Integration

NVIDIA has consistently positioned itself at the leading edge of the artificial intelligence (AI) landscape, primarily through its innovative computing solutions and rapid product development cycles. At the Consumer Electronics Show (CES) 2026, NVIDIA unveiled its Vera Rubin AI series, which is now in full production, featuring six newly engineered chips, inclusive of Vera CPUs and Rubin GPUs. In a recent interview with Korean media, NVIDIA’s CEO, Jensen Huang, expressed that this year’s GTC promises to reveal products that have “never been seen before.”

We’ve prepared several new chips the world has never seen before. Nothing is easy because all technologies are at their limits.

– NVIDIA’s CEO

This bold assertion from Jensen reflects his well-documented history of delivering on ambitious technology claims. Although specific details regarding the new chip were not disclosed, the reference to “never seen before”innovations leads us to speculate on possible candidates. One likely contender could be an advanced version of the Rubin lineup, such as the Rubin CPX. Alternatively, NVIDIA might introduce its next-generation Feynman chips, which are anticipated to be groundbreaking.

NVIDIA Blackwell chips with H100 T1 and B100 labels — Image Credits: NVIDIA

In the current computing landscape, NVIDIA is navigating a highly dynamic market where demand fluctuates regularly. For instance, while the Hopper and Blackwell architectures prioritized pre-training tasks, the recent Grace Blackwell Ultra and Vera Rubin platforms have shifted the focus towards inference, highlighting concerns over latency and memory bandwidth. There are indications that NVIDIA is also investigating more advanced SRAM-centric designs, potentially involving the integration of Logic Processing Units (LPUs) via 3D stacking, although this remains speculative.

Moreover, Jensen emphasized the critical role of strategic acquisitions and collaborations in maintaining NVIDIA’s competitive edge within the AI sector:

NVIDIA has great partners and amazing startups, and we are investing across the entire AI stack. AI is not just a model; it’s an entire industry encompassing energy, semiconductors, data centers, the cloud, and the applications built on top of it.

For those interested, NVIDIA’s GTC keynote is scheduled for March 15th in San Jose, California, where the discourse will center on the next phase of the AI infrastructure competition.

Source & Images