01.ai advances pre-IPO financing, targets $1B-plus HK listing by 2027

01.ai, the AI startup founded by Kai-Fu Lee, is advancing a pre-IPO financing round ahead of a planned Hong Kong listing in 2027, after pivoting from building foundational large language models to customizing existing open-weight systems.

"01.ai is in pursuit of pooling funds in Hong Kong to finance research and support the costly infrastructure required for AI services," Lee, who previously led Google's China operations, said. The company is unwinding its offshore holding structure, a prerequisite for smoother overseas listing procedures.

Founded in 2023, 01.ai reached a valuation of over $1 billion with backing from investors including Alibaba Cloud. The company now employs about 240 people — lean by industry standards — and no longer positions itself as a maker of proprietary LLMs after DeepSeek's release of open weights disrupted the economics of training AI models. Lee aims to complete the financing round around the time 01.ai releases its first annual results, though he has not elaborated on the fundraising target or potential investors.

The strategic reset means 01.ai now focuses on fine-tuning or customizing existing Chinese open-weight models — such as DeepSeek, Alibaba's Qianwen, or Z.AI's GLM — wrapping them in software that organizes customer data and deploying AI agents on top. This application-layer approach requires significantly less capital than training foundational models, which can cost tens of millions of dollars per training run for frontier systems.

The DeepSeek Disruption Reshapes AI Economics

DeepSeek's release of open-weight models fundamentally altered the competitive landscape for Chinese AI startups. The move made high-performance LLM capabilities available at a fraction of the cost of training proprietary models, forcing companies like 01.ai to reassess their strategies. Rather than competing on model architecture, 01.ai now differentiates through its software layer that integrates customer data with existing models and deploys AI agents for enterprise use cases.

The pivot mirrors a broader industry trend. Chinese AI companies that once raced to build the largest foundational models are increasingly shifting toward application-layer services, where margins depend on deployment efficiency rather than raw model performance. DeepSeek's open-weight strategy effectively commoditized the model layer, compressing the market opportunity for proprietary LLM builders and accelerating the shift toward fine-tuning and agent-based solutions. Companies like Baidu and Tencent have similarly adjusted their AI strategies to emphasize application-layer deployment over pure model development.

Hong Kong Listing Path and Investor Implications

01.ai's planned Hong Kong listing comes as the city's exchange pushes to attract more tech and AI listings, positioning itself as a primary destination for mainland Chinese AI companies seeking public market access. The company's unwinding of its offshore holding structure is a standard procedural step for mainland Chinese companies pursuing overseas listings, designed to align with regulatory requirements under Beijing's oversight framework.

For investors, the key question is whether 01.ai can generate sustainable revenue from its fine-tuning and AI agent business without the moat of a proprietary foundational model. The company's lean headcount of 240 employees suggests capital efficiency, but the AI services market is increasingly crowded, with major cloud providers like Alibaba and Tencent offering similar capabilities through their own platforms. 01.ai's valuation of over $1 billion from its last funding round reflects early investor confidence, but the success of its pre-IPO financing will signal whether institutional investors see value in the application-layer AI strategy as the company prepares for its 2027 listing.

This article is for informational purposes only and does not constitute investment advice.