ModelBest, a four-year-old Chinese AI start-up specializing in compact, edge-based artificial intelligence models, has started its mandatory pre-IPO tutoring process as it aims for a public listing in mainland China. The company focuses on AI models optimized for local device deployment, navigating chip shortages by leveraging domestic computing power.
- Pre-IPO tutoring initiated with Citic Securities for mainland China listing.
- Focus on lightweight AI models optimized for domestic chips amid US chip shortages.
- Collaboration with Huawei, Samsung, SAIC Motor, and Intel boosts deployment prospects.
What happened
ModelBest has officially launched the pre-IPO tutoring process required for companies planning an initial public offering on mainland Chinese stock exchanges. This step, overseen by Citic Securities, involves a structured preparation program covering corporate governance, financial audits, risk management, and disclosure practices. The process will last at least three months before a formal IPO application may be submitted.
Founded in August 2022, ModelBest specializes in edge AI models—compact artificial intelligence systems that operate locally on devices like smartphones and vehicles. This strategy enables the company to sidestep global chip shortages by tailoring AI to run on Chinese-made processors. The company recently released MiniCPM5-1B, a competitive 1-billion-parameter model built using Huawei’s Ascend chips, and has secured partnerships with prominent firms in consumer electronics and automotive sectors.
Why it matters
ModelBest’s approach to AI development is significant because it addresses key challenges facing China’s tech industry, especially its reliance on foreign semiconductor technology. By focusing on lightweight, efficient AI models compatible with domestic chips, ModelBest exemplifies a growing effort to localize AI capabilities amid geopolitical pressures and supply chain constraints.
The company’s successful funding rounds, which have raised over 5 billion yuan this year and pushed its valuation to 20 billion yuan, reflect strong investor confidence in its small-model edge AI concept. ModelBest’s collaborations with major technology and automotive players underscore the commercial viability and industry demand for AI systems that operate without relying on cloud infrastructure, reducing latency and increasing privacy.
What to watch next
Market observers will be closely monitoring ModelBest’s IPO timeline and fundraising target, which remain undisclosed. The successful completion of the pre-IPO tutoring stage and subsequent listing would mark a milestone for edge AI development in China and potentially inspire similar companies to pursue local listings.
Additionally, how ModelBest’s edge AI solutions penetrate sectors like smart vehicles, robotics, and consumer electronics will be a key indicator of broader adoption trends. Competing global efforts, such as Meta’s Muse Glimmer and Apple’s on-device models, highlight a fast-evolving landscape where local AI computing power on devices is seen as a competitive advantage for privacy, performance, and cost control.