As the global AI race spotlights ever‑larger models, several Chinese startups are staking their bets on mobile AI — compact models engineered to run entirely on smartphones and laptops. Their pitch: faster processing, stronger data privacy, and lower operating costs.
Unpacking the bet against scale
This trend runs counter to the move toward massive, sometimes trillion‑parameter models that demand energy‑hungry data centers. By prioritizing on‑device AI, these companies aim to push intelligence to the endpoint and avoid constant round trips to the cloud.
Why on‑device AI appeals to investors and operators
Proponents of the local approach highlight three main benefits: reduced latency (inference happens immediately on the device), improved data privacy (sensitive information stays on the terminal), and lower infrastructure costs (less reliance on expensive data centers). These selling points have attracted funding for solutions tailored to the power and memory constraints of mobile hardware.
A case in point: ModelBest and the flow of capital
China’s edge AI market is highly active. In a WeChat post by investor Shanghai Mejoy Capital, Beijing‑based ModelBest — a startup focused on on‑device AI — reportedly raised more than ¥5 billion (roughly $738 million) over the past year. The post says a recent Series C+ pushed the company’s valuation to ¥20 billion. ModelBest did not disclose the exact amount of its latest round and did not immediately respond to a request for comment.
