DeepSeek Releases DeepInfer Engine with Native Support for Nvidia and Huawei Chips
DeepSeek has released DeepInfer, the inference framework powering its large-scale model services, with support for both Nvidia GPUs and Huawei Ascend processors.
Dick Weisinger · October 24, 2026 · 2 min read · Source: TechNode · Issue #136
East Asian Technology Intelligence
Tech news from Japan, China, Korea & Taiwan — translated, contextualized, and delivered for Western readers.
Free. Unsubscribe anytime.
This story ran in Issue #136, alongside three other stories.
AI & Machine Learning
What happened
DeepSeek has released DeepInfer, the inference framework powering its large-scale model services, with support for both Nvidia GPUs and Huawei Ascend processors. The release includes distributed inference, speculative decoding, and structured output capabilities, alongside deepseek-recipe for handling API requests and input-output processing. The framework currently supports DeepSeek-V4.1-Flash-1017, with further optimizations planned for subsequent model iterations.
Key details
DeepSeek released the DeepInfer inference framework on October 20.
The framework supports both Nvidia GPUs and Huawei Ascend processors.
DeepInfer currently supports DeepSeek-V4.1-Flash-1017. What remains unclear
Independent benchmark performance figures on Huawei Ascend hardware
Adoption rates among enterprise customers using non-Nvidia configurations
Full hardware compatibility list beyond Ascend processors and Nvidia GPUs
Source: TechNode · source published October 21, 2026
The inclusion of Huawei Ascend support alongside Nvidia hardware illustrates how domestic Chinese AI infrastructure development continues to adapt to export controls by building cross-hardware compatibility into core serving stacks.
Source
DeepSeek open-sources inference framework with support for Nvidia and Huawei chips