Kai-Fu Lee's 01.AI seeks pre-IPO funding for a 2027 Hong Kong listing

Why it matters

01.AI's IPO plan turns Lee's enterprise pivot into a public financial test. Investors will need to see how quickly reported orders become revenue and whether hands-on deployments can produce repeatable margins.

Kai-Fu Lee's 01.AI is seeking a pre-IPO financing round before pursuing a Hong Kong listing in 2027, Bloomberg reported on July 20.

Lee plans to close the round around the publication of 01.AI's first annual results. He has not disclosed a fundraising target, prospective valuation, investor lineup, offer size or precise IPO timetable. Both transactions remain plans: 01.AI has yet to close the financing or file for a listing.

The proposed IPO would add another chapter to a career that has moved between research, Big Tech, venture investing and company building. Lee earned his doctorate at Carnegie Mellon in 1988 under Raj Reddy, where his work produced an early large-vocabulary, speaker-independent continuous speech-recognition system. He later worked at Apple and Silicon Graphics before joining Microsoft Research in 1998, then served as president of Google China and founded Sinovation Ventures.

That background made Lee an unusually credible recruiter and fundraiser when he founded 01.AI in Beijing in 2023. It also gave him direct access to the executives who now sit at the center of 01.AI's commercial strategy.

A listing built on a different 01.AI

Lee initially presented 01.AI as a Chinese answer to OpenAI. In a November 2023 interview, he argued that China needed domestic large language models because OpenAI and Google did not offer their products in the market. 01.AI released the bilingual Yi-34B model and reached a reported $1 billion valuation after financing from Lee's Sinovation Ventures, Alibaba Cloud and undisclosed investors.

01.AI's route to the public markets now rests on enterprise deployments rather than an escalating contest to train the largest general-purpose model. Its website emphasizes consulting, private deployment, industry models and forward-deployed engineers who work inside customer operations. The approach resembles Palantir's service-heavy model, with software shaped around each customer's data, workflows and decision structure.

01.AI launched its TrueNorth decision platform alongside Boss AI, Investor AI and TopSales AI. Boss AI is designed to show executives operational risks and track commitments. Investor AI analyzes investment materials, financials and due-diligence records. TopSales AI connects CRM records, meetings, contracts, delivery data and customer communications to identify sales risks and recommend next actions.

The products reveal the economic thesis behind the pivot. Lee is selling systems tied to revenue, margins, investment decisions and executive accountability, areas where customers may tolerate longer deployments and larger contracts. 01.AI also markets sovereign AI projects to governments and local partners outside the United States and China.

This model requires substantial implementation work. 01.AI says its forward-deployed engineers connect data, construct business ontologies, build prototypes and validate deployments alongside customers. In a media discussion after the TrueNorth launch, Lee said the products are not yet ready for out-of-the-box use, require support from forward-deployed engineers and are not affordable for small and medium-sized businesses.

Orders are running ahead of recognized revenue

Lee said in the 36Kr interview that 01.AI booked RMB500 million in 2025 orders and RMB250 million in audited revenue, and that orders topped RMB1.5 billion in the first five months of 2026. Those figures came from Lee's company-wide anniversary letter and have not been independently established through public financial statements.

The gap between orders and recognized revenue will be central to any IPO review. Lee said orders are billed in installments and revenue recognition takes one to three years. An order therefore represents contracted work rather than revenue that 01.AI can immediately record.

Lee has said 01.AI aims to produce a profitable quarter before the IPO. The timing matters because a deployment-heavy business can show rapid bookings while carrying the engineering, sales and collection costs for years. Publishing annual results for the first time should provide investors with a clearer view of gross margins, operating expenses, cash use and the rate at which reported orders become revenue.

The planned pre-IPO round gives Lee an opportunity to finance that delivery cycle before public investors examine it. The undisclosed terms also leave the most important fundraising questions unanswered. 01.AI has not identified a lead investor, target amount or expected valuation, and its current valuation cannot be inferred from the $1 billion figure reported in 2023.

Lee remains part of the sales engine

01.AI's enterprise strategy draws directly on Lee's personal network. He told 36Kr that 01.AI is targeting roughly 2,000 to 3,000 large enterprises worldwide and has contacted 500 to 600 of them. Lee also said that two weeks earlier he had addressed nearly 200 chief executives through a speech and met 30 of them one-on-one.

That access helps 01.AI reach buyers who can authorize large, multi-year transformation projects. It also raises a scaling question: how much of 01.AI's pipeline depends on executives agreeing to meet Lee personally?

Lee has addressed that dependence directly. He told 36Kr that a growing proportion of projects involving top corporate leaders were being won by the team rather than by him. He described his own role as making initial contact, converting some orders and maintaining relationships, and said 01.AI was embedding those capabilities in its team and products. Turning those introductions into a repeatable sales process will matter as much as the software itself.

The 2027 listing plan puts a deadline on that work. Lee has already repositioned 01.AI once, moving from a capital-intensive foundation-model race toward enterprise systems designed to affect financial results. The next phase requires evidence that 01.AI can convert its founder's access, reported order book and hands-on deployments into durable revenue and profitable operations.