A planned funding round at Hangzhou, China-based hedge fund and unexpected AI juggernaut DeepSeek has probably been cancelled. According to a very delicately worded Bloomberg report, this comes just on the heels of the supposed leak of a long, off-the-cuff conversation with the company’s reclusive founder and CEO, Liang Wenfeng, that went viral. The conversation in question would read as extremely dull to most people, but went viral for being more than a bit impolitic if you’re an extremely AI-brained person.
The actual viral quotes are sort of lost in the Backrooms, if you will, and can’t be officially confirmed to have happened at all. They were apparently posted by a news publication owned by Chinese tech conglomerate Tencent, but there’s now a dead link in older articles and blog posts about this incident. The English language version of the transcript purportedly survives as an un-authenticated PDF posted to Github, but don’t blindly trust that link.
So part of what makes the delay of the funding round so intriguing to DeepSeek watchers is that it’s making analysts, like Kyle Chan of Brookings, more confident that the interview was real in the first place.
I was a bit hesitant to believe the leaked DeepSeek investor call. Seemed almost too good to be true. This makes me a bit less skeptical.
If it's real, Liang Wenfeng was incredibly detailed about working with Huawei, chip production bottlenecks, CUDA moat being broken, etc.
[https://t.co/UgYPPAissf]— Kyle Chan (@kyleichan)
[July 25, 2026]
Anyway, the reason the quotes attributed to Liang went viral—and this is my own analysis—seems to be that they were a little more hedged and careful about the rise of China as an AI superpower than Chinese AI talk tends to be, and sounded closer to U.S. conventional wisdom. Essentially, Liang reportedly said China’s lack of compute resources and reliance on Nvidia explains the gap in model capability between the U.S. and China, and that lack of compute leads to a talent brain drain, *even though Chinese talent is perfectly fine overall. *
However, Liang reportedly implied in the interview that Nvidia was the agent of its own downfall. Chinese companies are being forced to find workarounds, largely via Huawei, that allow them to bypass Nvidia’s hardware and software ecosystem, and Liang’s supposed remarks sounded confident that a breakthrough would come in a year or so—a bit of a bold prediction, but not exactly a surprising thing for a Chinese AI CEO to say.
Nonetheless, according to Bloomberg, anonymous, knowledgable people say Liang—who is generally described as shy and nerdy—is experiencing “frustration” due to the online reaction to what he apparently said in the leaked quotes, which according to Bloomberg were given to investors last month in the lead-up to an earlier funding round. So DeepSeek will reportedly not be signing the paperwork for its second funding round, which apparently had backers ready to go.* For now*. Bloomberg also writes that its anonymous informants say DeepSeek “might choose to resume the deal process at a later date.”
The meteoric rise of DeepSeek has made Liang very rich very quickly. Forbes currently rates him as the 50th richest person in the world, well above Anthropic’s Dario Amodei, who is ranked 186. DeepSeek models have a reputation for efficiency and rock-bottom prices compared to the frontier models. The company has reportedly begun the process of making an initial public offering. The actual planned date for that IPO is reportedly sometime before the end of this year.