Huawei Technologies will launch two new artificial-intelligence semiconductors in 2027, rotating chairman David Wang said, as the Chinese company builds its AI computing business and seeks to challenge US chipmaker Nvidia. Huawei plans to launch the 960DT in the first quarter of 2027 and the Ascend 960PR in the third quarter, Wang said. [1]
Wang said at a company event in Shanghai that Huawei's UnifiedBus technology would be a key link in its next generation of large AI systems, intended to help chips share information and work together more efficiently. Most advanced AI programmes require more processing power than a single chip can provide, the report added. [1]
Huawei has developed 11 semiconductors based on UnifiedBus technology for use in its large systems, Wang said, while its largest linked systems, which the company calls superclusters, can support up to 1 million AI processors. The company has shipped more than 1,000 smaller linked systems, called supernodes, to more than 370 customers, according to Wang. [1] [1]
Wang did not say how many chips could be connected within a single supernode, name customers, or give a breakdown of shipments. He also said Huawei's AI chip ecosystem had 5,270 monthly active developers. [1] [1]
The reported plans come amid US-China competition over the chips, data centres and software that underpin AI. Washington has imposed export controls restricting China's access to certain advanced computing chips and manufacturing equipment, while Chinese companies and policymakers have emphasised domestic alternatives, making the ability to connect chips increasingly important. The report noted Huawei still faces Nvidia, whose hardware and software tools are widely used by developers worldwide. [1] [1]
Sources & limitations
Research coverage is limited to a single readable secondary source, The Indian Express, republishing a Reuters report; no original document, company transcript or independent corroboration was captured.
The commissioning question asked how much of Huawei's supercluster claim rests on UnifiedBus interconnect versus process-node parity with Nvidia; the captured source reports the one-million-processor figure without technical detail, independent test results or Nvidia comparison, so that assessment could not be made.
The captured source does not report the earlier-draft framing about a 'breakthrough' in UnifiedBus, process-node parity, or the effectiveness of export controls beyond the general policy context quoted above.
Wang's product timelines are forward-looking plans for 2027, not completed or independently verified launches.
The source does not specify how many chips a supernode connects, which customers received shipments, or a shipment breakdown, so those gaps are disclosed rather than filled.
No original document was captured. Available evidence consists of publisher reporting.
Only one readable source was captured; independent corroboration is not established.
Some candidate sources were inaccessible or outside the supported research domains; consult the capture log.
[1] indianexpress.com · China’s Huawei sets 2027 launch for new AI chips as it targets Nvidia | Technology News - The Indian Express ↗
17/9/2026 · Publisher reporting
View the preserved source record →