China's Xiaomi to Invest Over USD8.7 Billion in AI in Next Three Years, CEO Says(Yicai) March 20 -- Xiaomi plans to spend more than CNY60 billion (USD8.7 billion) on artificial intelligence over the next three years, according to the chief executive of the Chinese electronics giant.
Xiaomi's first smartphone AI agent, the Xiaomi MiClaw, has started closed beta tests, Lei Jun, who is also founder and chairman of the Beijing-based company, said at a press conference yesterday. The tool is equipped with the MiMo large models and deeply integrated with the firm's operating system and "Human x Car x Home" smart ecosystem, he added.
Xiaomi unveiled three MiMo large models: the flagship base model MiMo-V2-Pro, the full-model Agent model MiMo-V2-Omni, and the text-to-audio model MiMo-V2-TTS. The MiMo-V2-Pro is a trillion-parameter-scale large model that supports a million-level context window, with its early internal test version, known as Hunter Alpha, sparking heated discussions on application programming interface platform OpenRouter.
"Xiaomi will likely make breakthroughs in technological innovation this year," Lei noted earlier this month, adding that the company has a good shot at achieving a grand convergence of self-developed chips, OS, and AI models on one terminal product this year.
Xiaomi will continue to hike investment in the research and development of core underlying technologies and plans to invest CNY200 billion (USD30 billion) to focus on chips, AI, OS, and other relevant areas over the next five years to maintain long-term competitiveness in high-end, intelligent, and green development paths, Lei pointed out.
In addition, Xiaomi launched its second-generation SU7 electric vehicle yesterday, priced at CNY219,900 (USD31,870) for the standard model, CNY249,900 for the Pro, and CNY303,900 for the Max, all CNY4,000 (USD580) cheaper than the first-gen.
Xiaomi has sold 381,000 first-gen SU7s, with 258,164 units sold last year, becoming the only sedan priced at over CNY200,000 to outsell Tesla's Model 3, according to Lei.
Editor: Martin Kadiev