DeepSeek pauses $1.5B round after leaked

DeepSeek has told prospective backers it is suspending a second funding round, sought at roughly 10 billion yuan (about $1.5 billion) at a pre-money valuation near 480 billion yuan (about $71 billion), well above the roughly $50 billion valuation from its first round, which closed at $7 billion in June. The pause follows viral social-media posts carrying comments attributed to founder Liang Wenfeng about China’s AI chip shortage and its gap with US labs, reportedly made during private investor meetings and leaked without his consent.

Here’s the part that needs a caveat: the specific numbers circulating, including claims that Liang said DeepSeek needed 200,000 Huawei chips but received a fraction of that, and that the company sits 12 to 18 months behind leading US labs while running on a small share of their compute, come from a transcript whose authenticity Bloomberg says it has not independently verified. DeepSeek had not responded to a request for comment on the transcript as of the reporting. What is confirmed is the pause itself and its timing, tied to Liang’s reported frustration at the leak rather than to any dispute over the deal terms; DeepSeek says talks could still resume.

What it means for operators

If you route any workload to DeepSeek’s models or are evaluating its API against other Chinese open-weight options, the fundraising story matters less than the compute-gap admission would, if confirmed: a chip-constrained DeepSeek is a business-continuity question, not just a competitive one. Wait for DeepSeek’s own statement or independently sourced reporting before treating specific compute figures as fact, but keep a fallback model in your routing config regardless, the same posture worth taking toward any lab whose funding or chip supply is in public dispute.