fuxi a0 gpu demoed at ICCAD 2025, built on Imagination Technologies dxtp derivative

xiang dixian showed the Fuxi A0 graphics card at the International Conference on Computer-Aided Design 2025 with demo units running a localized Lenovo New Vision Z-engine 3D platform. the card is reported to be built on a 5 nm process node and is said to have entered mass production.

imagination technologies introduced its DXTP GPU IP earlier this year as an efficiency-oriented architecture for smartphone and other power-constrained platforms. months later a derivative of that architecture has appeared in discrete form aimed at workstation use. xiang dixian exhibited the Fuxi A0 graphics card at the International Conference on Computer-Aided Design 2025, with functional demo units on display and hooked up in working systems. a localized version of the Lenovo New Vision Z-engine 3D platform was shown running on the demo hardware during the late November event.

reports from the region state that the Fuxi A0 is built on a 5 nm process node and has entered a mass production phase. xiang dixian positions the A0 model as optimized for rendering and gaming, and company representatives claimed that the GPU can run Black Myth: Wukong at pleasing frame rates. coverage referenced an additional variant, B0, which is marketed toward training and inference workloads rather than gaming. that B0 variant is said to include an onboard NPU offering ‘FP8 capabilities.’

the brief public demonstration and regional reports emphasize the vendor’s claim that the latest Fuxi GPUs can compete with familiar mainstream cards in ray tracing and upscaling (super resolution) performance. taken together, the public demo, the claimed mass production status, and the existence of a training-focused B0 variant suggest xiang dixian is targeting both workstation graphics and machine learning inference markets with a family derived from a mobile-focused architecture.

50

Impact Score

Introducing Mistral 3: open artificial intelligence models

Mistral 3 is a family of open, multimodal and multilingual Artificial Intelligence models that includes three Ministral edge models and a sparse Mistral Large 3 trained with 41B active and 675B total parameters, released under the Apache 2.0 license.

NVIDIA and Mistral Artificial Intelligence partner to accelerate new family of open models

NVIDIA and Mistral Artificial Intelligence announced a partnership to optimize the Mistral 3 family of open-source multilingual, multimodal models across NVIDIA supercomputing and edge platforms. The collaboration highlights Mistral Large 3, a mixture-of-experts model designed to improve efficiency and accuracy for enterprise artificial intelligence deployments starting Tuesday, Dec. 2.

Contact Us

Got questions? Use the form to contact us.

Contact Form

Clicking next sends a verification code to your email. After verifying, you can enter your message.