Alibaba, Ant Group join PyTorch Foundation

Alibaba Cloud, Cambricon and Ant Group used PyTorch Conference China in Shanghai on September 8 to formalize a deeper role inside the PyTorch Foundation, the Linux Foundation-hosted group behind the open-source AI framework and related projects. PyTorch said Alibaba Cloud and Cambricon joined as Platinum members and Ant Group joined as a Gold member, while Huawei took part as an existing member in conference keynotes on the open AI stack. The move matters because PyTorch Foundation membership is not just branding. The foundation says it supports PyTorch and a wider set of open-source AI projects including vLLM, Ray, DeepSpeed, Safetensors and Helion, giving members a venue to work on training, inference and deployment across shared infrastructure. ### Who actually joined, and at what level? (pytorch.org) PyTorch said Alibaba Cloud and Cambricon entered as Platinum members, the foundation’s top published membership tier in this announcement, while Ant Group joined as a Gold member. Huawei was not listed as a new entrant in the September 8 announcement; the Linux Foundation release described it as a long-standing member participating in the Shanghai event. (community.linuxfoundation.org) Cambricon’s separate membership announcement said Platinum status includes one seat on the PyTorch Foundation Governing Board. That gives at least some of the new entrants a formal role in policy, mission and technical direction inside the foundation’s governance structure. ### Why did this happen at a conference in Shanghai? (pytorch.org) Shanghai was the setting because PyTorch Conference China 2026 was co-located with KubeCon + CloudNativeCon and the OpenInfra Summit at the Shanghai International Convention Center on September 7-9. The organizers said the event was designed to bring together the cloud native, infrastructure and AI communities around practical deployment and shared industry use cases. (pytorch.org) The Linux Foundation said representatives from the three new member companies and Huawei were delivering keynotes on advancing the open AI stack, with topics spanning hardware, models and infrastructure. That framing puts the announcement inside a broader push to connect model development with the systems used to train and run those models in production. (linuxfoundation.org) ### What does the PyTorch Foundation say these companies bring? Alibaba Cloud’s earlier membership post said the company brings a global AI cloud network and the Qwen family of open-weight models, which PyTorch described as widely adopted by developers since 2023. That gives the foundation a member with both cloud infrastructure and model distribution reach. (linuxfoundation.org) Cambricon’s membership post emphasized AI chips and said the company would contribute from the hardware side of the stack. Ant Group’s inclusion as a Gold member adds another large Chinese technology company with experience in large-scale production systems, according to the September 8 PyTorch and Linux Foundation announcements. (pytorch.org) ### Why does Huawei show up if it was not a new member? Huawei appeared in the announcement because the September 8 release was framed around a joint appearance by Alibaba Cloud, Ant Group, Cambricon and Huawei in Shanghai, not because all four joined that day. The PyTorch and Linux Foundation statements both said the new memberships were Alibaba Cloud, Cambricon and Ant Group, and described Huawei as an existing participant in the foundation’s work and conference program. (pytorch.org) That distinction matters for readers trying to parse the headline. Three companies joined on September 8; Huawei was part of the event lineup and the broader member coalition already in place. ### What happens next inside the foundation? (pytorch.org) PyTorch Conference China continues in Shanghai through September 9, according to the event page and Linux Foundation schedule. The next concrete step is the conference program itself: keynotes and technical sessions from Alibaba Cloud, Ant Group, Cambricon and Huawei tied to hardware, models, infrastructure and production deployment. (pytorch.org) (linuxfoundation.org)