Alibaba Cloud plans to use fewer Western chips, to boost its already huge AI margins
Alibaba Cloud says rising margins on its AI services mean new hardware can pay for itself faster than previously expected, and it plans to lean more heavily on its own chips rather than commercially bought Western ones to widen those margins further. The shift matters because it signals both growing confidence in AI as Alibaba's main growth driver, at a time when its core e-commerce business is slowing, and a deepening move by a major Chinese cloud provider away from Western silicon.
CFO Toby Xu said servers run for five years and typically pay back costs within three, but AI-specific gear could shorten that payback period to 2.5 years, or even two years with reduced capital spending. Alibaba spent $10 billion on infrastructure last quarter, 75% more than a year earlier, while over 650 external customers now use cloud resources built on its own chips, still far behind AWS's 120,000-plus Graviton users. AI-driven cloud revenue grew 45% to $7.14 billion, well behind e-commerce's $30.34 billion (up just 4%) and still trailing AWS, Microsoft and Google, though Alibaba expects cloud revenue to hit around $10 billion next quarter; growth abroad remains constrained by Western government restrictions on Chinese cloud services.
- Alibaba Cloud aims to boost AI margins by cutting reliance on Western chips
- AI hardware payback period could shrink from three years to 2.5 or two
- Cloud AI revenue up 45% to $7.14bn, still trailing US rivals
New here? Start with this
Alibaba Cloud is the part of Alibaba that rents computing power and storage to businesses. It is investing heavily in equipment for artificial intelligence (AI), while the company’s much larger online shopping business is growing more slowly.
Alibaba’s finance chief, Toby Xu, says AI services can earn enough to repay the cost of specialised servers sooner than ordinary servers. The company is also developing its own chips, which could reduce its reliance on chips bought from Western suppliers.
The shift matters because Alibaba is betting that AI can become a stronger source of growth, while joining a wider effort in China to build technology with less dependence on Western suppliers. Its cloud business is expanding, but remains smaller than leading US rivals, and restrictions on Chinese cloud services limit its reach abroad.
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Alibaba's shift toward proprietary chips reflects sound business practice that every major technology company pursues, developing hardware optimised for specific workloads to improve efficiency, reduce costs, and shorten infrastructure payback periods. Over 650 customers already rely on Alibaba's chips successfully, demonstrating the approach's viability. Building domestic technological capacity is a legitimate strategic objective that allows companies to reduce vulnerability to supply disruptions and foreign restrictions whilst advancing innovation.
The case against
The broader fragmentation of Chinese and Western technology ecosystems—of which this strategy is one piece—creates serious concerns about global cooperation on artificial intelligence governance and safety. As these systems become incompatible through mutual supply-chain restrictions and strategic decoupling, coordinating on crucial challenges like preventing harmful AI applications, establishing security standards, and ensuring interoperability becomes technically much harder. This technological bifurcation, though perhaps rational for individual companies, may ultimately leave all parties worse off by preventing the kind of international coordination needed to safely develop powerful AI systems.