Nvidia's Next-Gen GPU: Codenamed Rubin, Debuting Late 2025

Nvidia, a name synonymous with cutting-edge graphics technology, is already setting its sights on the future. Reports suggest that the tech giant's next-next-gen GPU is codenamed Rubin, with a planned debut slated for late 2025. However, gamers eagerly awaiting the new tech will need to exercise patience, as gaming variants are not expected until late 2026.

Engineers' Insights: The Secrets Behind Future Tech

Delving into the world of senior CPU or GPU engineers can be a fascinating journey. These individuals, far more than just marketers, hold the keys to a realm where innovation is always a few steps ahead. When asked about upcoming products, engineers often need to reminisce about past projects, as they are already deeply involved in creating the next-next generation.

Currently, details about Nvidia's next-gen Blackwell gaming GPUs are scarce. Even if we were to inquire with senior engineers, they would likely need to think back to Blackwell, which is likely in the final stages of development before its anticipated release later this year.

Vera Rubin: The Inspiration Behind the Name

The next generation of Nvidia GPUs is rumoured to be named Rubin, in honour of Vera Rubin, a trailblazing astronomer. Rubin's pioneering work shed light on the properties of galaxies and provided compelling evidence for the existence of dark matter.

Ming-Chi Kuo's Insights: What to Expect from R100 GPU

Ming-Chi Kuo, a TF Securities analyst, has shared valuable insights regarding Nvidia's R100 GPU. According to Kuo, the R100 GPU is set to enter production in the fourth quarter of 2025, with full system solutions expected to follow in the first half of 2026. Nvidia's naming convention suggests that the R100 will be an AI-focused GPU, rather than a gaming variant.

Cutting-Edge Technology: TSMC's N3 Node and HBM4 Memory

The Rubin GPU is said to be built using TSMC's N3 node and will feature HBM4 memory. This design approach emphasises power efficiency, a critical consideration as high-end AI GPUs are on track to exceed the kilowatt range per card. Efforts to enhance power efficiency are crucial, particularly as data centres brimming with these systems present increasingly complex management challenges.

The Future of Computing: Big Chips and High Price Tags

Kuo's report mentions a 4x reticle design and Chip-On-Wafer-On-Substrate-L (CoWoS-L) technology. Depending on the number of chiplets the future R100 GPU incorporates, it could result in an exceptionally large chip. Nvidia could potentially command a price tag of $50,000 for such a chip, a figure that would not be out of the realm of possibility.

The Road Ahead: Speculations on Rubin Gaming GPUs

As for when we might see Rubin gaming GPUs hit the market, only time will tell. If we follow the release cadence of Blackwell, a late 2026 debut seems plausible, with mainstream cards likely to follow in 2027. In the fast-paced world of computing, this timeline feels like an eternity, highlighting the rapid evolution of technology in this space.

In conclusion, Nvidia's Rubin GPU represents a significant leap forward in graphics technology. With a strong emphasis on power efficiency and cutting-edge design, the Rubin GPU is poised to redefine the possibilities of AI and gaming graphics.