Alibaba Doubles Down on AI Ambitions with Major Platform and Hardware Push

Alibaba US01609W1027

Alibaba Group is significantly scaling up its artificial intelligence initiatives, committing substantial resources to both software platforms and the underlying hardware infrastructure. This aggressive dual-track strategy solidifies the Chinese tech giant's position as a dominant force in the nation's rapidly expanding AI sector. Recent developments, from a new corporate platform operating system to potential high-volume chip purchases, underscore the depth of this commitment.

The financial performance of Alibaba's Cloud Intelligence Group highlights the strategic importance of its AI investments. For the quarter ending September 2025, revenue from the cloud segment surged 34% to reach 39.8 billion yuan. Within this, products related to artificial intelligence continued their explosive trajectory, registering triple-digit year-over-year growth for another consecutive period.

This growth is fueled by massive capital expenditure. Over the past four fiscal quarters, Alibaba has invested approximately 120 billion yuan into AI and cloud infrastructure. Company CEO Eddie Wu has suggested that the original three-year investment target of 380 billion yuan "may have been set too low," a clear signal of the intense demand for AI computing power the company is experiencing.

DingTalk Unveils Dedicated AI Agent Ecosystem

At a recent event in Hangzhou, Alibaba's enterprise communication platform DingTalk introduced a specialized operating system built for AI agents, dubbed "Agent OS." The platform simultaneously revealed DingTalk Real, a suite of hardware designed to act as a physical terminal for these digital agents. According to CEO Chen Hang, all future AI agents on the DingTalk platform will be built upon this foundational Agent OS.

The rollout includes industry-specific AI products. For the education sector, DingTalk demonstrated an automated homework grading machine capable of evaluating 100 assignments within four minutes. For corporate clients, the platform is developing AI agents to automate functions such as recruitment, travel expense reporting, and marketing design. One highlighted agent, an AI travel assistant, is intended to autonomously plan itineraries, compare prices, and submit expense reports.

Strategic Chip Procurement Amid Intensifying Competition

Alibaba is reportedly considering a substantial purchase of 40,000 to 50,000 MI308 AI accelerators from Advanced Micro Devices (AMD). This chip, specifically developed for the Chinese market, has already received necessary U.S. export licenses. The MI308's appeal is multi-faceted: it boasts 192 GB of HBM3 memory, double the capacity of Nvidia's China-specific H20 chip (96 GB), and carries a price tag around $12,000 per unit—roughly 15% cheaper than the Nvidia alternative. Furthermore, the AMD chip has historically attracted less regulatory scrutiny.

This potential procurement occurs against a backdrop of fierce competition for AI computing resources in China. Nvidia has informed its Chinese customers that it will begin supplying H200 chips to the market by mid-February, with an initial volume estimated between 40,000 and 80,000 units. Alibaba's exploration of a dual-source chip procurement strategy emphasizes its determination to secure ample computing capacity for its expanding AI operations, ensuring it is not reliant on a single supplier.

The company's comprehensive approach—building proprietary software platforms while aggressively securing the hardware to run them—demonstrates a full-stack ambition to capture value across China's AI ecosystem.