Alibaba Group Holding Ltd has stepped up its in-house AI semiconductor program, aiming to reduce reliance on third-party suppliers amid global memory-chip shortages. The company’s Semiconductor Research and Development Division has reportedly doubled its engineering headcount over the past six months, with plans to tape out its next-generation AI accelerator by late 2024. Alibaba’s initiative follows delays in shipments of Nvidia H200 chips to Chinese customers and is designed to support its cloud-computing arm, Alibaba Cloud, which accounts for roughly 25% of the group’s quarterly revenue. Executives have set an internal target for domestically produced chips to comprise at least 30% of Alibaba Cloud’s AI workload capacity by the end of 2025.

In the latest trading session, Alibaba shares closed at $163.65, marking a 2.81% decline from the prior day— a steeper fall than the 1.6% drop in the benchmark Hang Seng Index. Trading volume on the Hong Kong and U.S. listings rose by 18% above the 30-day average, suggesting elevated investor concern. Analysts attribute the sell-off to profit-taking after a 15% rally over the past two months and renewed worries about regulatory scrutiny in Beijing. Consensus estimates for Alibaba’s next quarterly revenue growth have been trimmed to 9.5% year-over-year, down from 11.2% three weeks ago.