Beijing-based ModelBest has reached a valuation of 20 billion yuan after closing its C+ Series funding round earlier this week, positioning itself at the forefront of China's fast-expanding on-device artificial intelligence sector. The milestone underscores a growing conviction among investors that lightweight, locally-run AI models — not cloud-dependent giants — could define the next phase of consumer technology.

The case for smaller AI

While trillion-parameter models continue to dominate headlines, a parallel race is intensifying around AI systems compact enough to run entirely on smartphones or laptops, without routing data through remote servers. Proponents argue this localised approach delivers faster processing, stronger data privacy, and meaningfully lower operating costs compared with cloud-based alternatives.

The shift reflects a pragmatic recognition that not every AI workload requires a sprawling data centre — and that latency, connectivity constraints, and privacy regulations can make on-device inference a superior choice for a wide range of applications.

Funding momentum builds

ModelBest raised more than 5 billion yuan (US$738 million) this year, according to a WeChat post by investor Shanghai Mejoy Capital. The company did not disclose the specific size of its latest funding tranche and did not immediately respond to a request for comment. The C+ round completion pushed the firm's cumulative valuation to the 20 billion yuan mark, according to the same post.

The competitive backdrop

ModelBest's rise comes amid a broader surge of activity in China's local AI and edge AI sector. Established technology players including Baidu, Alibaba, and Huawei Technologies — whose Ascend chip line targets on-device inference — are all competing for position in this space, as are global names such as Apple, whose Apple Intelligence platform pursues a similar on-device strategy. Benchmarking platforms such as Artificial Analysis and emerging frameworks like PrismML are increasingly being used to evaluate these compact models. Start-ups such as StepFun, also based in Shanghai, are adding further competitive pressure.

Why it matters

The on-device AI segment represents a structural challenge to cloud hyperscalers, whose revenue models depend on inference workloads being processed remotely. If compact models reach performance parity with cloud counterparts for common tasks, the economics of AI deployment could shift substantially. For hardware makers and chipmakers serving the mobile ecosystem, the opportunity is equally significant.

What's next

Investors and industry observers will be watching whether ModelBest discloses further details of its C+ round and how it deploys fresh capital — particularly whether it accelerates hardware partnerships or expands its model portfolio. The broader edge AI race in China is expected to intensify through the second half of 2026, with device makers under growing pressure to differentiate on native AI capabilities.