Skip to content

AI & Automation

Analysis

Huawei launches Ascend 960DT AI chip in Q1 2027, doubles performance

Huawei has moved up the launch of its Ascend 960DT AI chip to Q1 2027 — from Q3 2027 — citing doubled performance and annual advancement. The chip will power new AI systems built on Huawei’s Peerium Computing Architecture, including the Atlas 950 SuperCluster, which supports up to 256,000 accelerator cards. Analyst Rui Ma noted a discrepancy between the earlier 15,488-chip Atlas 960 SuperPoD target and the newly announced 4,096-chip system. The announcement precedes the September 24 meeting between U.S. President Trump and Chinese President Xi Jinping and reflects Huawei’s push amid U.S. semiconductor restrictions.

Original source: techcrunch.com

Huawei launches Ascend 960DT AI chip in Q1 2027, doubles performance

Ask SCI.AI about this story

📋 本文要点

Huawei has moved up the launch of its Ascend 960DT AI chip to Q1 2027 — from Q3 2027 — citing doubled performance and annual advancement. The chip will power new AI systems built on Huawei’s Peerium Computing Architecture, including the Atlas 950 SuperCluster, which supports up to 256,000 accelerator cards. Analyst Rui Ma noted a discrepancy between the earlier 15,488-chip Atlas 960 SuperPoD target and the newly announced 4,096-chip system. The announcement precedes the September 24 meeting between U.S. President Trump and Chinese President Xi Jinping and reflects Huawei’s push amid U.S. semiconductor restrictions.

According to techcrunch.com, Huawei has accelerated the launch of its Ascend 960DT AI chip to Q1 2027, moving it up from its originally planned release in Q3 2027. The announcement was made by David Wang, Huawei’s rotating and acting chairman, at the company’s Huawei Connect conference on Thursday.

Accelerated timeline and performance claims

A Huawei spokesperson confirmed to TechCrunch that the Ascend 960DT is now expected to be ready in Q1 2027, and emphasized a consistent annual advancement in capability: “The Ascend 960DT is expected to be ready in Q1 2027. Ascend 960 chips are launching ahead of schedule, doubling performance and advancing year by year,” the spokesperson said.

The revised schedule places the chip’s debut just one week before the scheduled meeting between U.S. President Trump and Chinese President Xi Jinping in Washington, DC on September 24. Huawei’s push comes amid tightening U.S. export controls on advanced semiconductor technology, yet the company continues to advance its domestic chip development agenda.

According to the report, Huawei’s new Peerium Computing Architecture — built around its UnifiedBus interconnect technology — underpins its strategy to integrate hundreds of thousands of AI chips into unified computing systems. The architecture supports both AI training and inference workloads.

System scale and architectural rollout

Eric Xu, Huawei’s rotating chairman, stated in the company’s official statement that the Peerium Computing Architecture is being deployed to build larger AI computing systems. The first implementations are the Atlas 950 SuperPoD and Atlas 950 SuperCluster.

The company says an Atlas 950 SuperCluster can connect up to 256,000 accelerator cards — a figure representing the largest-scale system specification cited in the source. This contrasts with earlier plans for the Atlas 960 SuperPoD, which had been described as scaling to 15,488 Ascend 960 chips, versus the newly announced system using 4,096 chips.

China tech analyst Rui Ma highlighted this discrepancy on X, noting: “The chip itself is coming WAY earlier, but the SuperPoD they announced is much smaller than what they originally laid out.” She added that U.S. restrictions are unlikely to halt China’s semiconductor progress, stating, “I think it’s futile to stop China’s development in semiconductors because the stakes for self-sufficiency are just too high at this point.”

Geopolitical context and AI development race

The timing of Huawei’s announcement coincides with intensifying U.S.-China competition in AI infrastructure. While President Trump has resisted calls to slow AI development over safety concerns — arguing the U.S. must preserve its lead — Eric Xu countered that China must accelerate AI development to catch up, according to the Financial Times. Xu argued that Chinese companies need to advance further before they can fully understand and address risks posed by increasingly powerful AI systems.

Source: TechCrunch

Compiled from international media by the SCI.AI editorial team.

Ask SCI.AI Finished reading? Continue with SCI.AI. Explore the related policy, route, company and historical context. Continue asking
Stay ahead of supply chain shifts

One email a week: the stories that matter and new risk events — in English, free.

Prefer RSS? Articles Risk events

China tungsten export curbs drive PCB drill bit prices up several-fold
AI & Automation

China tungsten export curbs drive PCB drill bit prices up several-fold

China's tungsten export controls have driven up prices for the micro drill bits used to bore holes in AI server PCBs several-fold, Nikkei Asia reported. Yet SMM estimates that 2026 tungsten consumption by PCB micro drills and WF6 comes to only about 3,800 tonnes of contained metal, roughly 4.6% of China's total. Comparing USGS data, China's Ministry of Commerce control notice, SMM price data and Sumitomo Electric papers points to three factors: the processing stage of controlled raw materials, the material quality demanded of very fine drills, and diverging price trends inside and outside China. China produced an estimated 67,000 tonnes of tungsten in 2025, a 78.8% world share.

Knapp opens Boston robotics and AI hub with 35 staff
AI & Automation

Knapp opens Boston robotics and AI hub with 35 staff

Knapp has opened a robotics and AI innovation hub in Somerville, Massachusetts, under a new entity, KNAPPRobotics.ai Inc. The Austrian warehouse automation provider said the unit will employ 30 to 35 people by early 2027 and will work closely with MIT on AI, computer vision and robotics. Graz, Austria remains the center for robotics development and industrial readiness, while Boston serves as an international innovation hub for specific industrial applications. The unit consolidates product management, development and implementation, and is built around physical AI systems overseen by the KNAPP Brain platform.

Opentrons raises $40M Series E for AI lab automation
AI & Automation

Opentrons raises $40M Series E for AI lab automation

Opentrons Labworks has closed a $40 million Series E round led by Avego, with BroadOak Capital Partners also investing. The laboratory automation company will use the capital to scale its robotics, software, and data infrastructure for AI-driven autonomous science. Opentrons says it has deployed more than 11,000 robots to labs in over 80 countries, including at 90 percent of the top 20 global pharmaceutical companies. CEO James Atwood said millions of scientists still spend much of their day on manual work robots can handle, and the company aims to return that time to discovery. The funding will support AI-driven experimental execution, faster protocol development, and expansion into new markets and geographies.

Welcome Back!

Login to your account below

Create New Account!

Fill the forms below to register

Retrieve your password

Please enter your username or email address to reset your password.

Scan to share via WeChat

Open WeChat and scan the QR code to share

QR Code

Add New Playlist