Chinese artificial intelligence startup DeepSeek is preparing for a potential IPO on Shanghai's STAR Market, appointing GL Ventures partner Yan Wentao as its first chief financial officer. Working with underwriter CITIC Securities, the company targets a 2026 launch for a 2027 listing. DeepSeek is seeking a pre-IPO funding round at a 500 billion RMB ($74 billion) valuation, supported by a rapid commercial ramp nearing $500 million in annual revenue and a highly disruptive low-cost model pricing strategy.
DeepSeek CFO appointment
- ▪Yan Wentao, born in 1991, has previously made investments in artificial intelligence companies, including DeepSeek rival MiniMax, while at GL Ventures
- ▪DeepSeek is appointing Yan Wentao, a partner at GL Ventures, as its first-ever chief financial officer to manage investor relations and financial controls
Shanghai STAR Market IPO
- ▪DeepSeek has appointed underwriters, including CITIC Securities, to handle pre-listing preparations for a potential IPO on Shanghai's STAR Market
- ▪The Shanghai Stock Exchange published guidelines in June 2026 governing the fifth set of listing criteria for large-model companies, requiring IPO candidates to show at least one widely used operating model
- ▪DeepSeek is targeting 2026 to start its listing process, with the goal of completing the Shanghai STAR Market IPO in 2027
DeepSeek valuation trajectory
- ▪DeepSeek is reportedly nearing $500 million in annual revenue, representing a significant commercial ramp from its origins as a privately funded research operation
- ▪DeepSeek completed a funding round in June 2026, raising approximately 50 billion RMB (roughly $7 billion) at a valuation of 52 billion RMB, with backing from Tencent and JD.com
- ▪DeepSeek is seeking a follow-on funding round targeting a valuation of approximately 500 billion RMB (or $74 billion) ahead of its potential IPO
Low-cost AI pricing strategy
- ▪A study by Juniper Research suggests that Chinese artificial intelligence models can be operated at costs up to 90% lower than popular alternatives in the United States
- ▪The DeepSeek-V4.1-Flash model achieved a cached-input rate of $0.003, compared to $0.40 for GPT-5.6 Sol and $0.50 for Claude Opus 5, according to VentureBeat
US-China AI competition
- ▪The 2026 AI Index published by Stanford suggests that top-performing artificial intelligence models in the United States held only a 2.7% performance lead over top Chinese models as of March 2026
- ▪Enterprises have begun adopting Chinese technology over Western alternatives due to cost, including Thomson Reuters adopting Alibaba's Qwen as a substitute for Claude to manage document reviews
Debatable claims
- ▪Western enterprises should adopt Chinese AI models to cut costs
- ▪Chinese AI models will overtake US models in overall capability
Story comments
Loading comments…