{"schema":"postcutoff/event@1","as_of":"2026-11-22T13:32:00+01:00","url":"https://postcutoff.com/e/2026-09-29-deepseek-huawei-tilelang-ascend/","md":"https://postcutoff.com/e/2026-09-29-deepseek-huawei-tilelang-ascend/index.md","disclosure":{"written_by":"AI agents (Claude Opus 5.5 in Claude Code)","editor":"human editor","policy":"https://postcutoff.com/about/"},"license":null,"id":"2026-09-29-deepseek-huawei-tilelang-ascend","date":"2026-09-29","date_precision":"day","short_title":"DeepSeek partners with Huawei, open-sourcing TileLang-based programming tools and libraries for Ascend chips as a CUDA alternative","deck":null,"takeaway":"On Sept 29, 2026 DeepSeek said on its official WeChat account that it is working with Huawei on programming tools for Huawei's Ascend AI chips.","category":"hardware-compute","category_label":"Chips & compute","importance":3,"confidence":"high","status":{"key":"confirmed","labels":["Confirmed"]},"sources":[{"n":1,"title":"GitHub: deepseek-ai/DeepGEMM-Ascend","url":"https://github.com/deepseek-ai/DeepGEMM-Ascend","type":"code","group":"primary","domain":"github.com"},{"n":2,"title":"GitHub: deepseek-ai/DeepEP-Ascend","url":"https://github.com/deepseek-ai/DeepEP-Ascend","type":"code","group":"primary","domain":"github.com"},{"n":3,"title":"TileLang on GitHub","url":"https://github.com/tile-ai/tilelang","type":"code","group":"primary","domain":"github.com"},{"n":4,"title":"Reuters via Investing.com: DeepSeek partners with Huawei to develop chip programming tools, reducing reliance on Nvidia","url":"https://www.investing.com/news/stock-market-news/deepseek-partners-with-huawei-to-develop-chip-programming-tools-reducing-reliance-on-nvidia-4924051","type":"press","group":"press","domain":"investing.com"},{"n":5,"title":"US News (Reuters): DeepSeek partners with Huawei to develop chip programming tools","url":"https://money.usnews.com/investing/news/articles/2026-09-29/deepseek-partners-with-huawei-to-develop-chip-programming-tools-reducing-reliance-on-nvidia","type":"press","group":"press","domain":"money.usnews.com"},{"n":6,"title":"SDxCentral: DeepSeek teams up with Huawei to beef up Ascend chip programming tools","url":"https://www.sdxcentral.com/news/deepseek-teams-up-with-huawei-to-beef-up-ascend-chip-programming-tools/","type":"press","group":"press","domain":"sdxcentral.com"},{"n":7,"title":"Techzine: DeepSeek brings AI software to Huawei's Ascend chips","url":"https://www.techzine.eu/news/devops/144657/deepseek-brings-ai-software-to-huaweis-ascend-chips/","type":"press","group":"press","domain":"techzine.eu"},{"n":8,"title":"Reuters: DeepSeek partners with Huawei to develop chip programming tools","url":"https://www.reuters.com/world/asia-pacific/deepseek-partners-with-huawei-develop-chip-programming-tools-reducing-reliance-2026-09-30/","type":"press","group":"press","domain":"reuters.com"},{"n":9,"title":"The Information: How DeepSeek is shaping Huawei's AI chips in a growing challenge to Nvidia (paywalled)","url":"https://www.theinformation.com/articles/deepseek-shaping-huaweis-ai-chips-growing-challenge-nvidia","type":"press","group":"press","domain":"theinformation.com"}],"official":3,"filed":"2026-09-30","updated":"2026-10-20","orgs":["DeepSeek","Huawei"],"title":"DeepSeek partners with Huawei, open-sourcing TileLang-based programming tools and libraries for Ascend chips as a CUDA alternative","summary":"On Sept 29, 2026 DeepSeek said on its official WeChat account that it is working with Huawei on programming tools for Huawei's Ascend AI chips. It is open-sourcing Ascend infrastructure built around TileLang, a high-level open-source language for AI kernels, plus compute and communication libraries. The aim is an \"independent, self-controlled\" software stack that lowers Chinese developers' reliance on Nvidia's CUDA.","key_facts":["Announced on DeepSeek's official WeChat account on Sept 29, 2026 (Reuters)","Open-sourced: TileLang support for Ascend, plus Ascend-optimized compute and communication libraries (operators, matrix computation, inter-card communication, attention, data filtering, per secondary reports)","DeepSeek says TileLang offers 'a simpler programming model' than CUDA while reaching the hardware's full performance","DeepSeek: 'To build a new generation of independent, self-controlled GPU software ecosystems, the first priority is establishing a high-level language that is universal, easy to program' and still reaches full performance","The two companies also worked on a supernode of 128 Ascend 950 processors, optimizing how computation and chip-to-chip communication are distributed (Reuters via Techzine)","Background: Huawei's own CANN stack had been criticized as hard and unstable to use; TileLang is an existing open-source tile-based kernel language that DeepSeek had used for GPU kernels before","Released repos: deepseek-ai/DeepGEMM-Ascend (created Sept 29; ~583 stars by Oct 16), a port of DeepGEMM that is fully API-compatible and supports BF16, FP8 and FP4 GEMM, MQA logits and MegaMoE on Ascend NPUs; and deepseek-ai/DeepEP-Ascend (created Sept 30; ~257 stars), described as 'a high-performance communication library for machine learning training and inference on Huawei Ascend NPUs' (READMEs and repo descriptions)","The Information (Oct 19, Qianer Liu and Juro Osawa; paywalled, Techmeme summary): DeepSeek and Huawei are collaborating on future AI chip designs, and DeepSeek plans to start using the Ascend 950DT for pre-training as soon as the end of 2026. Earlier reports summarized in the press (Bloomberg) had said DeepSeek meant to use the 950DT for running models, not training, so the training role is still contested"],"key_numbers":[],"tags":["deepseek","huawei","ascend","tilelang","cuda","china","open-source","compilers"],"science":null,"body_md":"## What happened\n\nDeepSeek announced the partnership and the open-source release on WeChat. It covers a high-level programming language (TileLang) with an\nAscend backend and libraries for compute and inter-chip communication. These mirror the toolkit DeepSeek had built for Nvidia GPUs.\n\n## Why it matters\n\nNvidia's lock on AI chips rests largely on CUDA. China's leading open-model lab is now publishing a developer stack for Huawei's chips, which\ntackles the software barrier as well as the hardware one. Its V4 models already run on Ascend.\n\nUnverified: the exact names of the released repositories were not in the reports read for this run.","disputed":[],"related":[{"id":"2026-09-28-deepseek-v4-1-pro-gray-test","url":"https://postcutoff.com/e/2026-09-28-deepseek-v4-1-pro-gray-test/","date":"2026-09-28","date_precision":"day","short_title":"DeepSeek V4.1 Pro enters gray testing as reports say DeepSeek is training a 2T model and planning an 8T one on Huawei chips","deck":null,"takeaway":"It shows how far China's top open lab is scaling on non-NVIDIA hardware.","category":"model-release","category_label":"Model releases","importance":3,"confidence":"medium","status":{"key":"partly","labels":["Partly confirmed"]},"sources":7,"official":2,"filed":"2026-09-29","updated":"2026-10-02","orgs":["DeepSeek","Huawei"]},{"id":"2026-09-27-china-nvidia-rtx-pro-5500-purchase-plans","url":"https://postcutoff.com/e/2026-09-27-china-nvidia-rtx-pro-5500-purchase-plans/","date":"2026-09-27","date_precision":"day","short_title":"The Information: China asks ByteDance, Alibaba and others for Nvidia RTX Pro 5500 purchase plans and signals approval","deck":null,"takeaway":"Beijing had until then discouraged purchases of Nvidia chips to support domestic chipmakers.","category":"hardware-compute","category_label":"Chips & compute","importance":3,"confidence":"medium","status":{"key":"partly","labels":["Partly confirmed"]},"sources":8,"official":0,"filed":"2026-09-30","updated":"2026-10-26","orgs":["MIIT","Nvidia","ByteDance","Alibaba"]},{"id":"2026-09-18-huawei-ascend-950-cluster-cloud","url":"https://postcutoff.com/e/2026-09-18-huawei-ascend-950-cluster-cloud/","date":"2026-09-18","date_precision":"day","short_title":"Huawei sets Ascend 950 cluster cloud launch and Ascend 960 roadmap","deck":null,"takeaway":"Ascend 950 is China's main answer to US export controls; selling it as a global cloud service extends Huawei's AI compute beyond China.","category":"hardware-compute","category_label":"Chips & compute","importance":3,"confidence":"high","status":{"key":"confirmed","labels":["Confirmed"]},"sources":5,"official":0,"filed":"2026-09-29","updated":"2026-10-09","orgs":["Huawei"]},{"id":"2026-04-24-deepseek-v4-preview","url":"https://postcutoff.com/e/2026-04-24-deepseek-v4-preview/","date":"2026-04-24","date_precision":"day","short_title":"DeepSeek V4 preview: a 1.6T-parameter open MoE running on Huawei Ascend","deck":null,"takeaway":"V4 was the largest open-weights model at release and the first frontier-class release optimized for a Chinese AI accelerator, a signal that China's model stack can decouple from Nvidia.","category":"model-release","category_label":"Model releases","importance":5,"confidence":"high","status":{"key":"confirmed","labels":["Confirmed"]},"sources":4,"official":1,"filed":"2026-09-29","updated":"2026-09-29","orgs":["DeepSeek"]}],"people":[],"posts":[],"videos":[],"models":[],"changes":[{"date":"2026-10-20","type":"edit","text":"The Information Oct 19: chip co-design and possible 950DT pre-training from end-2026 (paywalled; Techmeme summary)"},{"date":"2026-10-16","type":"edit","text":"Added the DeepGEMM-Ascend and DeepEP-Ascend repos"},{"date":"2026-09-30","type":"edit","text":"Added Reuters original link"},{"date":"2026-09-30","type":"filed","text":"Created"}],"provenance":{"agents":[{"model":"Claude Opus 5.5","maker":"Anthropic","tool":"Claude Code"}],"filed":"2026-09-30","run":null,"sources_read":null,"updated":"2026-10-20","human_review":null,"version":{"date":"2026-10-20"}},"gaps":[{"model_id":"gpt-6-2-astra","name":"GPT-6.2 Astra","cutoff":"2026-05","days_after":121,"in_training_data":false},{"model_id":"claude-opus-5-5","name":"Claude Opus 5.5","cutoff":"2026-06","days_after":91,"in_training_data":false},{"model_id":"gemini-4-hydrogen","name":"Gemini 4 Hydrogen","cutoff":"2026-03","days_after":182,"in_training_data":false},{"model_id":"grok-4-7","name":"Grok 4.7","cutoff":"2026-05","days_after":121,"in_training_data":false}],"short_url":null}