Resolution criteria

This market will resolve based on the total number of H100-equivalent chips manufactured by Huawei in the calendar year 2026. For the purpose of this market, one "H100 equivalent" is defined by the FLOPs capacity of an NVIDIA H100 GPU.

Per the user definition, both the 950DT and PR models will be counted as 0.5 H100 equivalents each, based on an estimated performance of 500 TFLOPS (BF16). Other Huawei AI chips will be calculated proportionally based on their rated BF16 TFLOPS capacity relative to the H100's benchmark.

The total count will be determined by official annual reports, confirmed production figures from reputable semiconductor industry analysts (such as TrendForce, IDC, or Gartner), or official statements from Huawei. If final 2026 data is not available, the market will resolve once authoritative third-party estimates for the year are published.

Background

Huawei’s semiconductor division, HiSilicon, has been scaling its production of AI accelerators, such as the Ascend series, in response to U.S. export restrictions limiting access to advanced NVIDIA hardware. The "H100 equivalent" metric is commonly used in industry discourse to compare the computational throughput of Chinese domestic alternatives to NVIDIA’s Hopper architecture. As of early 2026, Huawei’s production capacity remains subject to constraints involving domestic advanced-node manufacturing (such as SMIC’s capabilities) and supply chain logistics.

This description was generated by AI. Review and verify everything here yourself. You can edit, replace, or delete any part of this description, including the resolution criteria. You do not need to trust the AI output.