DeepSeek has crossed a $1 billion annualized revenue run rate and is on track to close a $6.88 billion second funding round at a $68.8 billion valuation by the end of October, chief executive Liang Wenfeng told investors on Sunday, according to reports on 24 September 2026.
Revenue more than doubled after price hikes
The Hangzhou-based lab was under $500 million in annualized revenue only a few months ago. It raised API prices between 2.3x and 4.5x last month without visible customer flight, letting the top line more than double even as capacity remains constrained by Chinese AI-chip availability. Revenue continues to be sourced almost entirely from paid API access to DeepSeek's V4-Pro and V4.1-Flash models.
Route to a Shanghai listing
Liang described the round as a bridge to a planned Shanghai Stock Exchange listing that would follow Unitree Robotics's own Shanghai IPO, where DeepSeek is already a strategic investor. Chinese state-linked funds have anchored the private rounds; the new tranche is understood to be led by Chinese sovereign vehicles and existing backers rather than Western venture firms.
Pressure on OpenAI and Anthropic pricing
DeepSeek's ability to raise API prices sharply and still grow throws further pressure on U.S. frontier labs — OpenAI just halved GPT-6 pricing and Anthropic released a faster, cheaper Claude Opus 5.5 — and shows that inference economics in China are moving on a different curve. If the $6.88 billion closes, it would be one of the largest AI-lab rounds ever from a non-U.S. investor base.
Reporting based on coverage from The Information, Dealroom and PYMNTS.
