The second software stack now has its own compiler. deepseek open-sourced its full ascend platform…
the second software stack now has its own compiler. deepseek open-sourced its full ascend platform kit on sept 30, tilelang plus five components mirroring the nvidia stack one-to-one, after porting every training operator, with huawei co-building the ascend 950 128-card supernode.
the export-control bet became a toolchain business.
Context
Verified date: DeepSeek open-sourced infrastructure components for Huawei's Ascend platform on Wed Sep 30, 2026. SCMP says it released six software modules that mirror its earlier open-source tools for Nvidia chips, and Tencent News lists TileLang plus DeepGEMM, DeepEP, TileKernels, FlashMLA and DeepSelect, each matching a component it had released for Nvidia.
Chip Dispatch says the stack was validated on a 128-chip Ascend 950 supernode, and Huawei open-sourced joint work on the Ascend 950 in the CANN community. Reuters reports the DeepSeek and Huawei partnership.
The Sep 30 date, TileLang, the one-to-one mirroring of Nvidia components and the 128-chip Ascend 950 supernode match the sources. SCMP counts six modules; Tencent lists TileLang and five named components, so the 'five components' in the post fits. That DeepSeek ported every training operator was not seen in the sources read, so unsupported here, not refuted. 'The export-control bet became a toolchain business' is the author's opinion.
Related work
- DeepSeek开源昇腾基础组件, Tencent News ↗Sep 30, 2026. Lists the components.
- SCMP ↗Sep 30, 2026.
- Reuters ↗Sep 30, 2026.
Watch next
- Read DeepSeek's repositories to check which Nvidia-side operators were actually ported.
Sources
- DeepSeek partners with Huawei to develop chip programming tools, Reutersreuters.com
- China's DeepSeek open-sources tools to help Huawei chips supplant Nvidia in AI, South China Morning Postscmp.com
- DeepSeek Open-Sources TileLang, a CUDA Alternative for Huawei Ascend AI Chips, Chip Dispatchchipdispatch.com
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 1 October 2026 at 02:18 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →