China to triple domestic Artificial Intelligence accelerator output with Huawei and SMIC

China plans to triple domestic Artificial Intelligence accelerator production by dedicating a fabrication plant to Huawei´s Ascend chips and bringing two more plants online, sources say. SMIC will assist as Huawei builds an end-to-end domestic silicon supply chain.

China is preparing a major scale-up of domestic Artificial Intelligence accelerator manufacturing to cut reliance on foreign computing, according to anonymous sources cited by the Financial Times. The plan reportedly dedicates an entire semiconductor fabrication facility to Huawei for production of its Ascend accelerators, with that plant scheduled to begin output by the end of the year. Two additional fabrication plants are expected to come online next year.

SMIC, China’s premier semiconductor manufacturer, has been supporting Huawei’s efforts, but the reporting indicates Huawei has moved to initiate its own independent production with SMIC’s assistance. Details on total wafer output and process node specifications for the dedicated fab were not stated. The expansion is presented as part of a broader push to significantly scale domestic manufacturing operations for accelerators used in Artificial Intelligence workloads.

Previous reporting cited in the article says Huawei, frustrated with limited SMIC capacity, has secured elements of the entire silicon production chain. That effort includes sourcing materials and chemicals, wafer fab equipment and chip-making equipment, and retaining chip design capabilities. A set of Huawei-backed companies forms the operation’s backbone: SiCarrier supplies optical and X-ray inspection tools, atomic force microscopes, alignment systems for metrology, gas-based and atomic layer deposition tools, plasma etchers, rapid thermal processors, and electrical testing platforms. SwaySure and Fujian Jinhua provide memory chips, Si´En and Pehgjin supply power chips, and PWX and PST handle logic.

78

Impact Score

NVIDIA and Doosan broaden physical Artificial Intelligence partnership

NVIDIA and Doosan Group are expanding work across robotics, autonomous equipment, power infrastructure and advanced materials. The partnership links NVIDIA accelerated computing platforms with Doosan businesses serving industrial automation, energy systems and data center hardware.

Chatbot liability suits test Artificial Intelligence safety law

A Florida lawsuit targeting ChatGPT’s maker signals a new product liability threat for Artificial Intelligence companies. The fight could turn on unsettled questions about platform immunity, speech protections, causation, and federal safety rules.

Canada pushes Artificial Intelligence sovereignty strategy

Canada has unveiled an Artificial Intelligence for All strategy focused on reducing reliance on foreign cloud and Artificial Intelligence providers. The plan mirrors the EU’s new sovereignty push and sets targets for adoption, infrastructure and jobs.

Contact Us

Got questions? Use the form to contact us.

Contact Form

Clicking next sends a verification code to your email. After verifying, you can enter your message.