Urgent.News

What's breaking now, across thousands of outlets.

AI

DeepSeek readies Huawei chip cluster for Inner Mongolia

DeepSeek is planning to install at least 160,000 Huawei Ascend 950DT artificial intelligence accelerators at a new data centre in Inner Mongolia, a deployment that could become one of the largest known clusters built around Huawei’s advanced AI processors. The Hangzhou-based AI developer is expected to use the chips primarily for inference, the process of running trained models and generating…

DeepSeek, a Hangzhou-based AI developer, is set to install a cluster of at least 160,000 Huawei Ascend 950DT artificial intelligence accelerators at a data centre in Inner Mongolia. This deployment could potentially become one of the largest clusters built around Huawei's advanced AI processors. The chips are primarily intended for inference, where trained models generate responses, rather than the more computationally intensive task of training frontier systems.

DeepSeek continues to use Nvidia accelerators for training, indicating that domestic processors still lag behind US hardware in certain workloads. The planned installation is part of a larger gigawatt-scale computing project in Ulanqab, situated about 350 kilometers northwest of Beijing. DeepSeek aims to bring at least part of the facility online by late 2027 or early 2028, depending on Huawei's ability to manufacture and deliver sufficient processors.

However, both DeepSeek and Huawei have not publicly confirmed the chip order, and the scale and timing remain subject to change due to constraints in high-end memory and other advanced components needed for production. The scale and timing of this deployment could significantly impact the availability of high-end computing resources in China, especially as AI labs, cloud providers, and internet groups compete for scarce accelerators capable of supporting larger models and growing inference demand.

Written by urgent.news from Arabian Post's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at thearabianpost.com →

More in AI

Hallucinations in Regulated Workflows: Catch Them at the Gateway with Bifrost

TL;DR In mission-critical sectors like healthcare, life sciences, and financial services, ungrounded model outputs breach regulatory mandates such as the EU AI Act, HIPAA, and NIST AI RMF guidelines.

  • Hallucinations in AI language models cause legal issues in regulated sectors.
  • Bifrost, an open-source AI gateway, detects and prevents fabricated statements.
  • Bifrost enforces guardrails at network level for immutable audit trails.

More from Saturday 5 September →