Mar 21, 2026 | Posted by Abdul-Rahman Oladimeji
Following the unveiling of its upcoming Language Processing Unit (LPU) chip at this week’s GTC conference, Nvidia has updated its data center roadmap through 2028 to include three hardware generations.
Maintaining its annual release cycle, Nvidia plans to roll out a new GPU architecture each year and, after its Christmas Eve licensing deal with Groq, is now also set to introduce a new LPU architecture annually.
The Nvidia Groq 3 LPU is slated for release in the second half of 2026, available in liquid-cooled LPX racks featuring 256 LPUs, 128GB of on-chip SRAM, and 640TBps of scale-up bandwidth, targeting low-latency AI inference workloads.
Nvidia will also launch its Vera Rubin platform later this year, combining Vera CPUs and Rubin GPUs, alongside additional components including the NVLink 6 Switch, ConnectX-9 SuperNIC, BlueField-4 DPU, and Spectrum-6 Ethernet Switch.
The company also plans to release its Kyber NVL144 rack-scale solution in 2027. Featuring 144 Rubin Ultra GPUs and connected via its NVLink 7 Switch, the system offers a 4x performance improvement over its Blackwell NVL72 system – also known as Oberon. Meanwhile, Nvidia’s second LPU, the Groq LP35 LPU, is also slated for release in the second half of 2027.
In 2025, following the Rubin Ultra announcement, Nvidia CEO Jensen Huang said that the transition to the platform would take “years of planning.”
“This isn’t like buying a laptop,” he said at the time. “We have to plan with the land and the power for data centers with engineering teams two to three years out, which is why I [am showing] the roadmap.”
According to Nvidia, the Feynman GPU will use die stacking and custom high-bandwidth memory to boost performance and expand memory bandwidth. The platform will also be the company’s first to feature NVLink switches with co-packaged optics.
Speaking at the 2026 conference, Huang said that demand for Nvidia products through 2027 had hit the $1 trillion mark. Nvidia CFO Colette Kress later confirmed that the number only related to Blackwell and Rubin products, along with associated networking, and did not include new products, like Groq LPUs and standalone CPUs.
"A trillion dollars is an enormous amount of infrastructure,” Huang said. “You have to have complete confidence that the trillion dollars you're putting down will be utilized, will be performant, will be incredibly cost-effective, and have [a] useful life for as long as you can see that infrastructure investment. [Nvidia] is the only infrastructure in the world that you could go anywhere in the world and build with complete confidence."
