Chinese artificial intelligence developer DeepSeek has postponed the launch of its latest AI model after failing to train it using chips from Huawei, the Financial Times reported Thursday.

The setback underscores the challenges China faces in its drive to reduce reliance on U.S. technology in advanced computing.

DeepSeek, one of China’s rising AI firms, had planned to unveil the model this month as part of efforts to compete with global leaders in generative AI. However, sources told the Financial Times that the system could not be successfully trained on Huawei’s processors, forcing the company to push back the release.

The delay highlights broader constraints facing Beijing’s tech ambitions amid U.S. export controls that have restricted China’s access to high-performance semiconductors from companies such as Nvidia. These chips are crucial for training large AI models, which require immense computational power.

Huawei, blacklisted by Washington in 2019 over national security concerns, has sought to position itself as a domestic alternative in the semiconductor and cloud computing space. Its chips are viewed by Chinese policymakers as a key element in countering U.S. sanctions and fostering technological self-sufficiency.

While Huawei has made progress in chip design and production, industry experts say its processors still lag behind top-tier U.S. offerings in performance and efficiency, especially for AI workloads. The DeepSeek setback is likely to fuel debate within China over how quickly its tech sector can close that gap.

Beijing has poured billions of dollars into developing an indigenous semiconductor supply chain, part of its broader “Made in China” strategy. But analysts warn that without equivalent hardware to match U.S. capabilities, Chinese AI companies may face further delays in rolling out competitive products for global markets.