Understand this story
On November 3, 2026, according to a report by Star Market Daily, multiple industry chain sources said that DeepSeek's next-generation large model, DeepSeek V5, plans to conduct pretraining based on Huawei Ascend chips; previously, Ascend chips were mainly used for inference tasks of DeepSeek large models. East Money reported the same day that after the release of DeepSeek-V4 in April 2026, Huawei said that its full series of Ascend SuperNode products could fully support the V4 series models; on September 30, DeepSeek open-sourced infrastructure components for the Ascend computing platform, covering TileLang compilation tools, DeepGEMM, FlashMLA, DeepEP, and others. The report also said that domestic computing power deployment is currently still concentrated on the inference side, and that the full-parameter post-training of DeepSeek-V4-Pro previously completed by an Ascend 910C cluster did not involve basic pretraining. There is currently no official response from DeepSeek or Huawei.
AI-generated from public reports
Latest update · Discussion
X.PIN joined this discussionDevelopments
1 confirmed developments
- Source:East Money
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Nov 4
X.PINRelated discussion
X.PIN joined this discussion
This post contributes to heat; it is not a formal report or confirmed fact.
@mark_kRelated discussion
@mark_k joined this discussion
This post contributes to heat; it is not a formal report or confirmed fact.
Semiconductor Industry Observer
DeepSeek V5 plans to pretrain on Ascend, as domestic AI chips enter a breakout year
Semiconductor Industry Observer reports that DeepSeek's next-generation large model, DeepSeek V5, plans to conduct pretraining based on Ascend chips, with core training compute pointing to Ascend 950 and the project timeline pointing to landing next year; previously, Ascend only played an inference role in DeepSeek's system. The article also reviews inference-side progress: third-party testing of the Iluvatar CoreX Tianyao 300 shows single-card end-to-end throughput reaching 1.05 times that of H200, and the MetaX Xiyun C600, with 144GB HBM3e memory, will enter mass production in May 2026. IDC data shows that in the first half of 2026, Nvidia's share of China's data center AI chip market fell to 34%, with Huawei Ascend following closely at 28%.
East Money
DeepSeek V5 Plans to Pretrain on Huawei Ascend Chips, the First Large Model Natively Trained on Ascend 950
DeepSeek's next-generation large model, DeepSeek V5, plans to conduct pretraining based on Huawei Ascend chips. Previously, Ascend chips were mainly used for inference tasks of DeepSeek large models. V5 will be its first large model natively trained on Ascend 950, with the project timeline pointing to a launch next year. Huaxi Securities noted that the driving force of domestic computing power has shifted from catching up on single-card performance to system-level engineering capability and supply guarantee capability, and manufacturers with system-level supply capabilities such as super-node complete machines will benefit.
Electronic Technology Application
DeepSeek V5 plans to pretrain on Huawei Ascend 950 chips
According to reports, DeepSeek's next-generation large model, DeepSeek V5, plans to conduct pretraining on Huawei Ascend chips. Previously, Ascend chips were mainly used for inference tasks in DeepSeek large models; V5 will be its first large model to natively rely on Ascend 950 for training. The Ascend 950 series is divided into two models, 950PR and 950DT, both based on the third-generation Da Vinci architecture. The 950DT is equipped with 144GB of self-developed HBM and a memory access bandwidth of 4TB/s, while the 950PR is equipped with 128GB of on-chip memory and FP4 compute power reaching 1784 TFLOPS. People familiar with the matter said V5 plans to use Ascend 950 as its core training compute, with the project timeline pointing to launch next year.
Nov 3
cnBetaRelated discussion
cnBeta joined this discussion
This post contributes to heat; it is not a formal report or confirmed fact.
Semiconductor Industry Review
Report: DeepSeek's Next-Generation Large Model V5 Plans to Use Ascend Chips for Pretraining
Reports suggest that DeepSeek's next-generation large model V5 will be pretrained on Huawei Ascend chips; previously, Ascend was mainly used for inference tasks on DeepSeek models. On September 30, DeepSeek open-sourced the TileLang compilation tool for the Ascend platform, computing libraries such as DeepGEMM, and the DeepEP communication library, corresponding one-to-one with NVIDIA platform components. The article also mentions that teams including Shenzhen Hetao College completed full-parameter post-training of the 1.6-trillion-parameter DeepSeek-V4-Pro on an Ascend 910C cluster, but this did not involve foundational pretraining.
East MoneySelected
DeepSeek V5 plans to pretrain on Ascend chips
According to sources in the industry chain, DeepSeek's next-generation large model, DeepSeek V5, plans to conduct pretraining based on Huawei Ascend chips; previously, Ascend was mainly used for inference tasks of DeepSeek models. The report mentions that after the release of DeepSeek-V4 in April 2026, Huawei stated that the full range of Ascend supernode products could fully support the V4 series models; on September 30, DeepSeek open-sourced infrastructure components for the Ascend computing platform, covering TileLang compilation tools, DeepGEMM, FlashMLA, DeepEP, and others. Currently, the implementation of domestic computing power remains concentrated on the inference side; the full-parameter post-training of DeepSeek-V4-Pro previously completed on the Ascend 910C cluster did not involve foundational pretraining.
East MoneyRelated discussion
East Money joined this discussion
This post contributes to heat; it is not a formal report or confirmed fact.
Phoenix TechnologyRelated discussion
Phoenix Technology joined this discussion
This post contributes to heat; it is not a formal report or confirmed fact.
Phoenix Technology
Report: DeepSeek V5 Plans to Use Huawei Ascend Chips for Pretraining
According to a report by the STAR Market Daily, multiple industry chain sources say that DeepSeek's next-generation large model, DeepSeek V5, plans to conduct pretraining based on Huawei Ascend chips. Previously, Ascend chips were mainly used for inference tasks of DeepSeek large models.
Cailian Press
DeepSeek V5 plans to conduct pre-training based on Huawei Ascend chips
According to multiple industry chain sources, DeepSeek's next-generation large model DeepSeek V5 plans to conduct pre-training based on Huawei Ascend chips; previously, Ascend was mainly used for inference tasks of DeepSeek models. The report outlined that around the Spring Festival of 2025, Ascend completed the large-scale deployment of DeepSeek-V3/R1 in Huawei Cloud's Gui'an Data Center; after the release of DeepSeek-V4 in April 2026, the full series of Ascend SuperNode products could already support the V4 series; and on September 30, DeepSeek also open-sourced infrastructure components such as TileLang, DeepGEMM, and DeepEP for the Ascend platform. The report also noted that the implementation of domestic computing power is currently still concentrated on the inference side, and the ecosystem migration cost on the training side is the variable that determines the market ceiling.
Cailian PressRelated discussion
Cailian Press joined this discussion
This post contributes to heat; it is not a formal report or confirmed fact.
Heat trend for this story
- Current
- 33
- Period peak
- 43 Nov 5 00:00
- 24-hour change
- +37%
Move the pointer or select the chart to inspect hourly heat. Use the left and right arrow keys with a keyboard.