Alibaba Group Holding is expected to report a 50 per cent revenue surge for its cloud and artificial intelligence unit for the September quarter, driven by strong returns on its aggressive AI investments, according to analysts. Analysts across several financial institutions projected revenue growth for the Chinese tech giant’s AI Cloud and Compute Services unit to have accelerated in the three months ended September 30, up from the 45 per cent growth seen in the preceding quarter. The unit’s profit margins are also expected to rise, signalling progress in Alibaba’s efforts to monetise its massive AI investments. Jefferies projected the segment’s profit margin to reach 12.3 per cent for the quarter, while Singapore-based brokerage Longbridge Securities forecast an EBITA (earnings before interest, taxes and amortisation) margin of about 15 per cent, up from 12 per cent in the June quarter. The earnings preview followed a strong showing in the June quarter, when Alibaba’s AI cloud segment – which began including the T-Head chip business that quarter – posted 48.4 billion yuan (US$7.2 billion) in revenue, its fastest growth rate in 22 quarters. Revenue for Alibaba’s AI Lab and Applications segment was projected to grow 20 per cent year on year, accelerating from a 16 per cent expansion in the previous quarter, Longbridge estimated. The unit – which is separate from the AI cloud business – encompasses the company’s AI model labs, the Qwen Consumer Business Group and workplace platform QwenWork. The brokerage expected losses for the unit to narrow to between 10 billion and 11 billion yuan for the period. Overall group performance was also expected to pick up. Jefferies estimated that Alibaba’s total September-quarter revenue would rise 9.6 per cent year on year, marking a slight improvement from the 9 per cent revenue growth recorded in the previous quarter. Despite Alibaba’s heavy capital expenditure, profits from the group’s cloud business are expected to offset the quarterly losses from the AI Lab and Applications segment by the end of December, according to Jefferies analysts Thomas Chong and Zoe Zong. The forecasts follow a slew of announcements at Alibaba’s annual Apsara Conference in September. During the event, the company debuted the Zhenwu V900 processor, billing it as “the most powerful AI chip in China today”, and teased plans to train an AI model with up to 10 trillion parameters. The Hangzhou-based company also outlined ambitions to scale its Alibaba Cloud global data centre capacity to more than 20 gigawatts by 2032. To fuel its AI ambitions, Alibaba issued HK$80 billion (US$10.2 billion) worth of new shares in August, with the firm planning to invest the proceeds to boost its full-stack capabilities, including by expanding its AI infrastructure. It was the company’s first share placement since its Hong Kong listing in 2019. Alibaba chairman Joe Tsai highlighted the company’s bet on full-stack AI capabilities at the annual Wave tech event in Italy on Wednesday. “We have distribution capability because of our very large e-commerce ecosystem, where we can apply AI to benefit our users and also our corporate clients,” he said, adding that AI would “be infused in every aspect of our business” over the next five years. Meanwhile, Alibaba’s core e-commerce business remains in a transition phase amid fierce domestic competition. Jefferies expected quarter-on-quarter losses for its quick-commerce unit to narrow to about 9.8 billion yuan. Alibaba shares dropped 1 per cent to HK$104.30 on Wednesday, amid a broader retreat of Chinese tech stocks in Hong Kong. The company is expected to report its September-quarter earnings in November. Alibaba owns the South China Morning Post.