Sovereignty incidents — hardware, export controls, supply chain
DeepSeek-V4 with day-0 Huawei Ascend adaptation (April 2026) [Added v0.7.1]
Claim
DeepSeek released V4 in preview on 24 April 2026 with open weights under the MIT licence (reported 1.6-trillion-parameter Mixture-of-Experts design with ~1M-token context; figures pending verification against the technical report). Huawei Ascend (950PR/950DT), Cambricon, Hygon, and Moore Threads completed day-0 adaptation at launch — an architectural decision to treat the domestic Chinese stack as a first-class target rather than a fallback, and the strongest evidence to date that a top-tier Chinese model no longer requires NVIDIA’s CUDA.
Sources
- Pandaily, “DeepSeek V4 Completes Full Adaptation to Huawei Ascend,” May 2026 ↗
- Huawei Central, “DeepSeek V4 and Huawei Ascend: a day 0 partnership” ↗
- “DeepSeek (chatbot),” Wikipedia (consolidating coverage) ↗
Referenced in
- Framework Part V (uncertainty resolved)
- AI-SAF-N Models pillar (L4 example)
- Software pillar
- State of Play §1
- §3.