Urgent.News

What's breaking now, across thousands of outlets.

AI

Huawei schedules two Ascend AI chips for 2027

Huawei Technologies has laid out plans to launch two new artificial intelligence processors in 2027, strengthening its push to supply the computing power China needs as US export controls restrict access to Nvidia’s most advanced chips. Rotating chairman David Wang said at Huawei Connect 2026 in Shanghai on Thursday that the company plans to release the 960DT in the first quarter of 2027 and the…

Huawei Technologies announced plans to release two new artificial intelligence processors in 2027, aiming to provide computing power for China despite US export controls restricting access to Nvidia's advanced chips. The company's rotating chairman, David Wang, revealed that the 960DT will launch in the first quarter of 2027, followed by the Ascend 960PR in the third quarter.

These accelerators are part of Huawei's strategy to build domestic alternatives for training and running AI models, reflecting a shift towards system-level performance over individual chip capabilities. Huawei unveiled the Ascend 960 SuperPoD and an upgraded UnifiedBus interconnect technology, capable of connecting up to 4,000 processors within a SuperPoD.

The company also mentioned developing the Atlas 960 SuperPoD and SuperCluster, each designed to combine over one million neural processing units. Huawei's current roadmap includes the Ascend 950DT, set to launch in the fourth quarter of 2026, featuring HiZQ 2.0 high-bandwidth memory for model training and inference workloads. US restrictions on China's access to advanced semiconductor manufacturing have given Huawei an opportunity to expand its domestic supplier base.

However, challenges remain, including shortages of high-bandwidth memory and the need to improve both hardware and software to reduce dependence on foreign computing infrastructure.

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

Also reported by 1 other outlet

Read the original at thearabianpost.com →

More in AI

The Vibe Coding Debt Trap: Why AI-Generated Code Breaks in Month 3

Originally published on tamiz.pro . The Illusion of Velocity vs. The Reality of Decay In the early days of integrating Large Language Models (LLMs) into the development workflow, the promise was…

  • AI-generated code introduces structural debt beyond variable naming issues.
  • LLMs optimize for next token, not long-term architectural integrity.
  • Month 3 Crisis reveals subtle inconsistencies, complex abstractions, and security vulnerabilities.

OpenAI reveals six new cases of AI misbehavior

US artificial intelligence giant OpenAI promised to more systemically report instances of its models going off track, while also publishing six new reports on previously undisclosed incidents of AI misbehavior.

More from Thursday 17 September →