DeepSeek fundraising pause tests AI investor demand in China

DeepSeek fundraising has been paused for now after online reports about founder Liang Wenfeng’s AI comments unsettled the Chinese lab’s talks.

The Chinese AI company has told some prospective investors in its second funding process that investment agreements will not be signed in the coming days, people familiar with the private discussions said. The delay comes only weeks after DeepSeek closed a first financing round in June that raised $7 billion.

Second-round documents stalled

DeepSeek delivered the message verbally to some would-be backers, according to people familiar with the matter, who asked not to be identified discussing private negotiations. The company may revive the process later, and the talks remain fluid.

It is not clear whether every prospective investor in the current round has received the same message. That leaves the process in a holding pattern rather than a formal cancellation, with timing now dependent on whether DeepSeek decides to reopen signing discussions.

The pause was driven in part by Liang’s frustration over online reports about comments he allegedly made to investors during the earlier financing process, the people said. Those reports circulated after the June round, when DeepSeek brought in major backers including Tencent Holdings Ltd. and Contemporary Amperex Technology Co. Ltd.

Liang comments hit sensitive nerve

The online material centered on what was described as a transcript of a meeting between Liang and unidentified parties. The authenticity of those posts has not been verified, and DeepSeek has not commented publicly on the transcript.

The reported remarks touched two of the most sensitive inputs in AI: access to Nvidia Corp. chips and the gap between China and the US in advanced AI capability. The source material said Liang was described as discussing China’s reliance on Nvidia hardware for AI development and the country’s continuing lag in AI sophistication versus the US.

DeepSeek did not immediately respond to an emailed request for comment on the transcript or the fundraising, according to the source material. That silence matters because the reported comments go directly to investor diligence: the durability of model progress, the availability of compute, and the political constraints around high-end chips.

$7 billion round reset expectations

The June financing gave DeepSeek a larger capital base and a more prominent roster of strategic investors. Tencent brings consumer internet scale and AI distribution reach, while Contemporary Amperex Technology is a major industrial name with deep exposure to China’s advanced manufacturing ecosystem.

The second round began only weeks after that deal closed, a short interval for a company already under close scrutiny. In a market where AI labs can burn capital quickly on talent, data, and compute, a new raise so soon after a $7 billion round signaled that investors were still trying to price DeepSeek’s position in China’s AI stack.

The pause changes that tempo. For DeepSeek, delayed agreements can slow momentum with new backers and force management to spend more time managing perception rather than product execution.

For the sector, the mechanism is straightforward: if investors believe Chinese AI champions face tighter compute ceilings or wider capability gaps than advertised, valuation discussions become more disciplined. If they instead treat the transcript dispute as a short-term communications problem, capital could return quickly to the strongest labs.

The global read-through is more about AI supply chains than one financing calendar. If chip access remains the binding constraint, money alone will not close the model-performance gap; if domestic alternatives improve, Chinese labs could regain bargaining power with investors and suppliers.

The immediate items to track are whether DeepSeek issues a statement, whether the second-round documents move back toward signing, and whether the disputed transcript is authenticated. The main uncertainty is whether the pause reflects a brief reputational flare-up or a deeper investor reassessment of DeepSeek’s compute strategy.