NVIDIA CEO Jensen Huang Affirms Blackwell Ultra & Vera Rubin AI Architectures Set to Debut at GTC 2025

NVIDIA has announced that its forthcoming architectures, Blackwell Ultra and Rubin AI, are progressing as planned, emphasizing that the next wave of AI technology will be broader and more impactful than ever before.

NVIDIA Set to Launch Groundbreaking AI Architectures in Unison

The next generation of NVIDIA’s AI offerings promises to be significantly more transformative than their predecessors. CEO Jensen Huang reiterated in a recent statement to analysts that both the Blackwell Ultra and Vera Rubin AI architectures are on track for their scheduled launches. Furthermore, he assured that the ongoing issues with the “GB200″yield rates would not derail NVIDIA’s annual release schedule.

In response to an inquiry from JPMorgan analyst Harlan Sur regarding NVIDIA’s simultaneous deployment of Blackwell Ultra and Vera Rubin, Huang stated:

Yes. Blackwell Ultra is second half. As you know, the first Blackwell was we had a hiccup that probably cost us a couple of months. We’re fully recovered, of course. The team did an amazing job recovering and all of our supply chain partners and just so many people helped us recover at the speed of light.

And so, now we’ve successfully ramped up production of Blackwell. But that doesn’t stop the next train. The next train is on an annual rhythm and Blackwell Ultra with new networking, new memories, and of course, new processors, and all of that is coming online.

– NVIDIA’s CEO

The anticipated Blackwell Ultra “B300″series is set to make a significant mark, with plans to boost power figures considerably. The upcoming GB300 AI server is expected to deliver an impressive thermal design power (TDP) of up to 1400W, a significant increase from the previous generation’s GB200, which operated at 1000W. Architectural enhancements suggest an estimated 1.4 times improvement in FP4 performance compared to earlier models, alongside an expanded memory capacity rising from 192 GB to 288 GB through the use of advanced 12-Hi stacks of HBM3E technology.

Positioned as a strategic link between the initial Blackwell series and the upcoming Vera Rubin architecture, Blackwell Ultra aims to enhance the available options for mainstream clients. As NVIDIA navigates yield rate challenges, the B300 series is perceived as a necessary reset—allowing the company to showcase its technological capabilities. Following the introduction of Blackwell Ultra, Vera Rubin is expected to bring groundbreaking innovations to the AI sector.

And the click after that is called Vera Rubin and all of our partners are getting up to speed on the transition of that and so preparing for that transition. And again, we’re going to provide a big, huge step-up.

– NVIDIA’s CEO

While specifics regarding the Rubin architecture remain scarce, its integration of HBM4 technology is anticipated to be critical in shaping the future of computing. Notably, both Blackwell Ultra and Rubin are likely to be unveiled at NVIDIA’s GTC 2025 conference in March, with their availability following closely behind. The industry is already preparing the necessary infrastructure to support these groundbreaking AI lineups.

Source&Images