Feynman (microarchitecture)
{{Short description|Nvidia microarchitecture}}
{{Infobox GPU microarchitecture
| name = Feynman
| image =
| caption =
| alt =
| launching = 2028
| discontinued =
| soldby = Nvidia
| designfirm = Nvidia
| manuf1 = TSMC
| process =
| codename =
| products-desktop1 =
| products-hedt1 =
| products-server1 =
| directx-version =
| direct3d-version =
| shadermodel-version =
| opencl-version =
| opengl-version =
| opengles-version =
| cuda-version =
| optix-version =
| mantle-api =
| vulkan-api =
| opengl-compute-version =
| cuda-compute-version =
| directcompute-version =
| compute =
| slowest =
| slow-unit =
| fastest =
| fast-unit =
| shader-clock =
| l0-cache =
| l1-cache =
| l2-cache =
| l3-cache =
| memory-support = HBM
| memory-clock =
| pcie-support =
| encode-codec =
| decode-codec =
| color-depth =
| encoders =
| display-outputs =
| predecessor = Rubin
| variant =
| successor =
}}
Feynman is a microarchitecture for GPUs by Nvidia announced at Nvidia GTC in 2025 by CEO Jensen Huang. It is named after theoretical physicist Richard Feynman. It will use a Vera CPU from the preceding generation microarchitecture Vera Rubin and is planned to be released in 2028.{{cite news|url=https://arstechnica.com/ai/2025/03/nvidia-announces-rubin-ultra-and-feynman-ai-chips-for-2027-and-2028/|title=Nvidia announces “Rubin Ultra” and “Feynman” AI chips for 2027 and 2028|author=Benj Edwards|date=March 18, 2025|work=Ars Technica}}{{cite news|title=While we despair of RTX 50-series supplies and wait on next-gen Rubin, Nvidia reveals its next-next GPU architecture will be known as Feynman and is due in 2028|url=https://www.pcgamer.com/hardware/processors/while-we-despair-of-rtx-50-series-supplies-and-wait-on-next-gen-rubin-nvidia-reveals-its-next-next-gpu-architecture-will-be-known-as-feynman-and-is-due-out-in-2028/|author=Jeremy Laird|date=March 18, 2025|work=PCGamer}}{{cite news|url=https://techcrunch.com/2025/03/18/nvidia-announces-new-gpus-at-gtc-2025-including-rubin/|title=Blackwell Ultra, Vera Rubin, and Feynman are Nvidia’s next GPUs|author=Kyle Wiggers|date=March 18, 2025|work=Tech Crunch}} Feynman will use High Bandwidth Memory (HBM).{{cite news|url=https://wccftech.com/nvidia-next-gen-feynman-gpu-updated-roadmap-arriving-2028-next-gen-hbm-memory/|title=NVIDIA Unveils Next-Gen Feynman GPU In Updated Roadmap, Arriving in 2028 With Next-Gen HBM Memory|author=Hassan Mujtaba|date=March 18, 2025|work=wccftech}} Nvidia is using its own Blackwell GPUs to accelerate the design of Feynman.{{cite news|url=https://www.eenewseurope.com/en/nvidia-accelerates-feynman-chip-design-manufacture-on-blackwell-gpu/|title=Nvidia accelerates Feynman chip design, manufacture on Blackwell GPU|author=Nick Flaherty|work=eenews|date=March 18, 2025}}