News
IndustryTSMC’s grip on advanced chipmaking will face competitionIndustryCadence joins European Chip Design Platform to support semiconductor startups and SMEsMaking chipsMarc Hijink on ASML’s Culture, Supply Chain, and GeopoliticsProcessorsIntel's post-Nova Lake CPUs take shape as Razor Lake desktop and laptop variants surfaceAI chipsHow Google’s New AI Chip Software Helps It Battle NvidiaIndiaMore IC Packaging Plants Emerge In IndiaIndustryUS Tightens China AI Chip Access as Nvidia, TSMC and Semiconductor Stocks SlideDesignAxelera AI moves into HPC with TitaniaMaking chipsChinese envoy says chip equipment export curbs threaten EU-China trade thawIndiaSEMICON India 2026 Day 2 Sees 25 MoUs, Strategic Collaborations Across Semiconductor Value ChainIndustryTSMC’s grip on advanced chipmaking will face competitionIndustryCadence joins European Chip Design Platform to support semiconductor startups and SMEsMaking chipsMarc Hijink on ASML’s Culture, Supply Chain, and GeopoliticsProcessorsIntel's post-Nova Lake CPUs take shape as Razor Lake desktop and laptop variants surfaceAI chipsHow Google’s New AI Chip Software Helps It Battle NvidiaIndiaMore IC Packaging Plants Emerge In IndiaIndustryUS Tightens China AI Chip Access as Nvidia, TSMC and Semiconductor Stocks SlideDesignAxelera AI moves into HPC with TitaniaMaking chipsChinese envoy says chip equipment export curbs threaten EU-China trade thawIndiaSEMICON India 2026 Day 2 Sees 25 MoUs, Strategic Collaborations Across Semiconductor Value Chain
EveryChip
Feynman

Feynman

Nvidia GPU microarchitecture planned for 2028 release

Feynman is a graphics processing unit microarchitecture designed by Nvidia. Announced in 2025, it is scheduled for launch in 2028 and will utilise High Bandwidth Memory.

Share on WhatsApp

  • Announced by Nvidia CEO Jensen Huang in 2025
  • Planned for release in 2028
  • Uses High Bandwidth Memory (HBM)
  • Named after physicist Richard Feynman
  • Designed using Nvidia Blackwell GPUs

What it is

Feynman is a microarchitecture for graphics processing units developed by Nvidia. The company announced the design at its GTC conference in 2025. It is named after the theoretical physicist Richard Feynman. This architecture is intended to succeed the Rubin microarchitecture. Nvidia plans to release products based on this design in 2028. The chips will support High Bandwidth Memory, which allows for faster data transfer rates than standard memory interfaces. This makes them suitable for demanding computational tasks.

How it came about

Nvidia CEO Jensen Huang revealed the Feynman microarchitecture during the Nvidia GTC event in 2025. The naming convention continues the tradition of honouring scientists, with this iteration paying tribute to Richard Feynman. The development process itself utilised previous generation hardware, as Nvidia employed its Blackwell GPUs to accelerate the design work. This approach highlights the iterative nature of modern chip development, where current tools build future ones.

Inside the chip

The Feynman architecture is built for graphics processing units. A key feature of its design is the integration of High Bandwidth Memory. This type of memory stacks dies vertically to provide wide data paths and high throughput. While specific details on core counts or clock speeds are not yet public, the focus on HBM suggests an emphasis on memory bandwidth for large datasets. The architecture follows the Rubin design in Nvidia's roadmap.

How it was made

TSMC is listed as the manufacturer for the Feynman microarchitecture. The specific process node or manufacturing technology has not been disclosed in the available information. The transistor count and physical packaging details are also currently unspecified. The design phase was aided by Nvidia's own Blackwell series of GPUs, which provided the computational power necessary to simulate and verify the new architecture before production.

Where it went

Feynman-based GPUs are expected to pair with the Rosa central processing unit. Rosa is a successor to the Vera CPU and was announced at GTC 2026. It is named after physicist Rosalyn Sussman Yalow. This pairing suggests that Feynman chips will be used in high-performance computing systems and servers where tight integration between the processor and graphics unit is beneficial. The use of HBM further points towards applications in artificial intelligence and data centres.

Versions and family

Feynman is positioned as the successor to the Rubin microarchitecture. No specific variants or sub-models have been detailed. The architecture is part of a broader ecosystem that includes the Rosa CPU.

Why it matters

As a future microarchitecture, Feynman represents the next step in Nvidia's GPU development timeline. Its announcement in 2025 sets expectations for the industry leading up to its 2028 launch. The use of Blackwell GPUs in its design demonstrates the increasing complexity of chip creation, requiring significant computational resources even during the engineering phase. The continued use of physicist names maintains brand consistency for Nvidia's high-end computing lines.

Questions people ask about Feynman

When will the Feynman microarchitecture be released?

Nvidia has planned the release of the Feynman microarchitecture for the year 2028. It was previously announced at the Nvidia GTC conference in 2025.

What type of memory does Feynman use?

The Feynman microarchitecture is designed to use High Bandwidth Memory, often abbreviated as HBM. This memory type is known for its high data transfer rates.

Who is Feynman named after?

The microarchitecture is named after Richard Feynman, who was a renowned theoretical physicist. This follows Nvidia's pattern of naming GPU architectures after famous scientists.

Which CPU will pair with Feynman GPUs?

Feynman GPUs are planned to be paired with the Rosa CPU. Rosa is the successor to the Vera CPU and was announced at GTC 2026, named after Rosalyn Sussman Yalow.

Good reading and sources

Where this page comes from

The facts on this page come from the Wikipedia article "Feynman (microarchitecture)", written by its contributors and used under the CC BY-SA 4.0 licence, and from its Wikidata record (CC0). EveryChip rewrote the text in its own words, with AI assistance, and compares every page with the article to make sure no passage is copied. It may be reused under the same CC BY-SA 4.0 licence with a link back.

It may contain mistakes, and figures change over time. Nothing here is investment or buying advice. If something is wrong, tell us. Last refreshed 2026-11-14.

Photo credits on this page

The small pictures above come from Wikimedia Commons under free licences. Some are cropped to fit.

The weekly digest

Every Monday: the chip of the day, new pages on EveryChip, and the conferences and anniversaries of the week ahead.