DeepSeek’s new AI model, V3.2-Exp, isn’t just a challenge to Alibaba’s AI division; it’s an indirect but significant pressure point on Alibaba’s core cloud computing business. As the lines between AI services and cloud hosting blur, DeepSeek’s aggressive pricing could have ripple effects across the entire cloud market.
Alibaba’s Qwen model is a key offering designed to drive customers to its cloud platform. When customers use Qwen, they are also consuming Alibaba Cloud resources, creating a powerful, vertically integrated ecosystem.
Now, DeepSeek enters with a highly efficient model and a 50% price cut. This creates a compelling reason for developers—even those already on Alibaba Cloud—to route their AI-specific tasks through DeepSeek’s API instead of using the native, more expensive Qwen model. This siphons off high-value AI workloads from Alibaba’s servers.
This dynamic forces Alibaba into a difficult position. To compete with DeepSeek’s AI pricing, it may have to lower the price of Qwen, potentially reducing the profitability of its AI services. Alternatively, failing to compete could mean losing AI customers and the associated cloud consumption revenue to other platforms where those customers might host their primary applications.
As an “intermediate step,” DeepSeek’s move is a clear warning. Its next-generation architecture will likely be even more integrated and competitive, posing a greater threat to the synergy between Alibaba’s AI and cloud offerings. This isn’t just a model-versus-model fight; it’s a strategic play in the broader cloud wars.