Enterprise AI Takes Center Stage
Kai Fu Lee's company 01.ai is pressing ahead with a Hong Kong IPO plan for 2027, and the Beijing-based firm is now working to close a pre IPO funding round before that listing happens. Lee told Bloomberg on the sidelines of the World AI Conference in Shanghai that the goal is to wrap up fundraising around the time the company shares its first-ever annual results, right after its fiscal year closes in December. It is a small update on paper, but it says something bigger about where Chinese AI investors are putting their trust these days, which is less on companies chasing the biggest model and more on those showing they can actually make money.
Strategy Before Scale
Back in 2023, 01.ai entered the market with big ambitions to build its own large language models, and it reached a valuation past US$1 billion with backing from names like Alibaba's cloud division. That story has changed quite a bit since then. Lee has more or less stepped away from the frontier model race, a decision that became easier to justify once DeepSeek's open-weight releases made training models from scratch a much tougher business case for most companies. These days, 01.ai builds on top of existing open-source and domestic models, including DeepSeek, Qwen from Alibaba, and GLM from Z.AI, then layers its own software over a client's messy internal data so AI agents can actually work with it. Lee calls this product Boss AI and often jokes that he wants 01.ai to become China's version of Palantir. The company is also untangling its offshore holding structure right now, a necessary step before any overseas listing can go smoothly.
Commercialization Becomes Differentiator
01.ai is not alone here. It belongs to a group of six companies known as China's AI Tigers, and most of them are facing the same pressure to show real revenue instead of just research promise. Z.AI and MiniMax have already gone public in Hong Kong this year, and StepFun, DeepSeek, and Moonshot are said to be lining up their own listings. Lee pointed out that only a handful of firms with truly massive balance sheets can still afford to build models from the ground up, leaving everyone else to figure out a different path forward. For 01.ai, that path is working. Around half of its revenue now comes from outside China, reaching parts of Asia, Europe, and South America, though Lee said the US remains off the table for now given ongoing wariness toward Chinese software there. Many clients also prefer keeping their AI systems on their own servers rather than in the cloud, even though that setup tends to be messier and more expensive to maintain.
Execution Drives AI Value
01.ai's journey toward its 2027 listing is a good reminder that AI companies are being judged differently now. With foundation models becoming easier to access, what really sets a company apart is how well it turns that access into something enterprises actually use every day. Every technology shift eventually becomes a business strategy story. InsightSphere helps leaders understand where innovation is creating measurable competitive advantage.