Leadership Turmoil and AI Ambitions Collide Ahead of Alibaba's Earnings

Alibaba faces key AI executive departures and investor scrutiny over costs as it prepares to report quarterly earnings with projected EPS decline and rising revenue.

As Alibaba prepares to release its quarterly financial results, the Chinese tech giant is navigating significant executive departures within its core artificial intelligence division. The surprise resignation of a key technical leader for its Qwen AI models triggered a sell-off, sending shares down as much as 5.3% in Hong Kong trading.

Quarterly Report Looms Amid Strategic Shifts

All eyes are on Thursday, March 19, when Alibaba will disclose earnings for the period ending December 2025. Market experts anticipate earnings per share of $1.72, a substantial decline from the $3.09 reported for the same quarter a year earlier. Revenue is projected to reach approximately $41.96 billion, representing an 8% year-over-year increase.

This report comes at a critical juncture. The company has fallen short of analyst expectations in three of the last four quarters. While cloud revenue surged 34% to RMB 39.8 billion in the most recent quarter, free cash flow swung dramatically from an inflow of RMB 13.7 billion to an outflow of RMB 21.8 billion. This reversal underscores the immense capital demands of its AI expansion, with investors closely watching cost management and the monetization progress of its cloud and AI services. The stock currently trades roughly 27% below its 52-week high.

Reshuffling at the Helm of Qwen AI

The recent volatility follows the unexpected departure of Junyang Lin, who played a central role in building the Qwen model family since April 2023. He stepped down just one day after the launch of the new Qwen-3.5 models. According to Reuters, two other senior figures have also left: Yu Bowen, who oversaw post-training for the Qwen models, exited in early March, and research scientist Hui Binyuan departed earlier in the year.

Alibaba's leadership moved swiftly to address the situation. CEO Eddie Wu confirmed Zhou Jingren as the new head of the Qwen unit. Furthermore, the company recruited Zhou Hao, a former senior research scientist at Google DeepMind with experience in Gemini research, to lead post-training research initiatives. Management has explicitly denied reports of a collective resignation by the core team or any imminent strategic shift away from its open-source approach.

Despite the executive turbulence, user adoption of Qwen products is accelerating rapidly. The application recorded roughly 203 million monthly active users in February, a massive jump from about 31 million in January. Since 2023, Alibaba has released more than 400 open-source Qwen models, which have collectively been downloaded over one billion times.

Building the Foundation: Chips and Infrastructure

Alongside the personnel changes, Alibaba continues to invest heavily in its underlying AI infrastructure. The company is constructing a major AI data center in Shanghai's Jinshan district, which will be powered by its internally developed Xuantie chips from Pingtouge Semiconductor. The firm reports it has already shipped hundreds of thousands of units of its Xuantie "Zhenwu" chip—a volume exceeding that of Cambricon Technologies. Its performance is said to be comparable to Nvidia's H20 processor.

In a complementary move, Alibaba Cloud has launched an AI-powered coding platform. This service provides developers with access to four Chinese open-source models under a single subscription, with a starting package priced at around six US dollars per month after the first introductory month.