Person
Xu Zhijun, known as Eric Xu, is one of Huawei's rotating chairmen, overseeing the Chinese tech giant's telecom, chip, and device businesses.
Huawei claims its Ascend chips have outsold Nvidia in China as of October 2026, but DeepSeek keeps its training work on Nvidia hardware. Unit sales and training workloads are separate measures, and sizing either company's China business needs both.
Reality
The Ascend 960 is dated 2027 and the 970 is dated 2028, but the only large order on the record is DeepSeek's 160,000 chips of the current generation. That fleet is forty pods of silicon Huawei ships today.
Perspective Coverage
8 publishersReality
DeepSeek open-sourced six modules for programming Huawei's Ascend 950, versions of tools it first built for Nvidia chips. The largest Ascend build tied to the release is DeepSeek's own plan for a data centre holding at least 160,000 of the chips.
Perspective Coverage
10 publishersReality
Wired's sources describe machines under evaluation for the most repeatable jobs on a data center floor, which turns technician headcount from a fixed operating input into something vendors now bid against.
Reality
The Atlas 960E fits 8 EFLOPS behind Huawei's own optical engines, and the Ascend 960DT arrives three quarters early, while rotating chairman Eric Xu says supply goes to Chinese customers first and dates China's catch-up with its own demand to 2030.
Perspective Coverage
4 publishersReality
Huawei moved its next AI chip three quarters forward to a first-quarter 2027 ship date, days before Xi Jinping's first White House visit in 11 years. Its chairman says Huawei may not make enough chips for domestic demand, so output remains the limit on Chinese AI compute.
Reality
Eric Xu told a Shanghai press conference that Chinese model providers should speed up until they hit the safety problems American labs keep disclosing, on the same day a US congressman asked those labs to help pace the frontier.
Reality
The 80% workload figure came from a technician rather than from Meta, and the arm that would deliver it still cannot match human speed, which makes this a pilot worth pricing rather than a substitution worth modelling.
Reality