AI large models are evolving from a business and product into the infrastructure a company runs on. Past software and hardware firms organised by product; AI was one feature among many. After models arrived, AI moved into product definition, user interaction and service distribution.
Globally, the shifts are similar. Google merged DeepMind and Brain. Meta raised AI priority around agents. In China, Tencent folded more AI into CSIG; ByteDance split Seed and Flow for model and application.
Xiaomi recently reorganised its XiaoAI and AI business. The near-decade-old product is moving from a standalone team to one backed by the base-model team, the cloud engineering team and the on-device OS team.
Per domestic media, XiaoAI’s tech splits into three: the MiMo base-model team led by Luo Fuli provides the underlying model; Luan Jian handles MiMo’s external engineering on XiaoAI; on-device capability moves to the phone and other OS teams.
MiMo must become the underlying capability
Before MiMo, AI lab, XiaoAI, phone OS, car and IoT teams each built AI in their own scene. MiMo did not immediately become the uniform choice. Teams can still pick models by effect, cost and launch cadence.
But MiMo cannot stay “optional.” As a long-term strategy, it must enter phone, car, home, wearable and robot scenarios, becoming a layer every line calls. Xiaomi’s business spans many terminals and systems; if each builds its own AI, you get a different assistant in the car, on the phone, at home. Experience fractures, and the “human-vehicle-home” ecosystem is impossible.
This reorganisation fixes that. Different terminals share a MiMo-based understanding, then each business schedules by its scene. Xiaomi’s strong internet culture, frequent cross-team talk and fast resource rallying suit this closed loop.
The base model becomes real infrastructure only if business teams feed scene data, user feedback and real call chains back into the model system. That binding brings pressure: MiMo must meet phone, car and IoT needs for stability, speed, cost and safety, and teams must accept deeper coordination.
For Xiaomi, burning cash on a base model looks costly with uncertain return today. But seen against how AI agents reshape entry points, services and value, this may be Xiaomi’s most strategically far-sighted AI investment.
Editor’s note: This is an adapted translation of the original LeiPhone report. It has been trimmed and restructured for readability for an international business audience. The full original (in Chinese) is at https://www.leiphone.com/latest/index/id/4755.