DeepSeek has told some prospective investors in its second fundraising round to pause, according to people familiar with the matter, after viral posts attributed to founder Liang Wenfeng circulated about US-China AI competition. The pause came quickly, with Liang verbally informing some would-be backers that they would not be signing investment agreements in the coming days as they had expected.

The deeper reason, per the same people, is founder frustration tied to the first financing round that closed in June and raised $7 billion for the AI lab. The suspension stems in part from Liang’s frustration over online reports about his comments to investors during that earlier deal, including a transcript of a meeting he held with unidentified parties. Bloomberg hasn’t verified the authenticity of those posts, but Chinese media including Yicai reported this week that Liang talked about reliance on Nvidia Corp. chips for AI development and China’s persistent lag in AI sophistication relative to the US.

That combination matters because fundraising in frontier AI is not just finance. It is trust, timing, and information control. When investor outreach goes public or is rumored to go public, negotiations can get messy fast. Even before anyone thinks about valuation mechanics, the counterparties have to ask a basic question: what did the founder say, to whom, and will those details change diligence, risk assumptions, or political exposure.

This pause is also happening right after DeepSeek’s record first round. DeepSeek closed its record first financing round mere weeks before the current talks, securing big names including Tencent Holdings Ltd. and Contemporary Amperex Technology Co. Ltd. (CATL). In the second round, the company was aiming to raise at least 10 billion yuan of additional funds in a follow-on deal, though the final amount could go higher depending on how many investors sign on. It had also targeted a pre-money valuation of at least 480 billion yuan, Bloomberg News previously reported, which marked an increase from the roughly $50 billion price tag it drew in its first round.

The pause is not necessarily a cancellation. The people say DeepSeek might resume the deal process at a later date, and negotiations are described as fluid. It is also unclear whether the company communicated its intentions to all prospective investors in the current deal. For decision-makers, that ambiguity is its own kind of risk: the fundraise could be temporarily stalled for some parties, while others continue to move forward, creating a split path where terms and timing differ by investor.

There is also the “what is the company actually optimizing for?” question, which DeepSeek has signaled it cares about beyond near-term commercialization. Bloomberg News has reported that DeepSeek’s senior management told potential investors the startup will prioritize groundbreaking AI research over short-term commercialization. Liang has pledged in at least one meeting with investors to keep developing open-source AI models while pursuing the broader goal of achieving artificial general intelligence, underscoring the company’s emphasis on advancing AI frontiers rather than monetization.

Those priorities connect to why investors care in the first place. DeepSeek developed a model last year that stunned the industry, demonstrating the ability to build a cutting-edge yet efficient platform with fewer computing resources. That mattered in a world where US export restrictions on advanced hardware squeeze China’s options. The breakthrough showed Chinese companies could compete with the best of Silicon Valley despite those constraints. Now DeepSeek is chasing more funds to support an ambitious expansion plan, including an increase in computing capacity. AI labs globally are striking deals to secure the data center infrastructure they need to train and run AI services, so capital is not just for research. It is for compute, storage, and the physical backbone of the model pipeline.

Finally, the capital strategy sits inside a longer runway that could reshape the Chinese tech landscape. DeepSeek has begun preparations for an initial public offering and may file as soon as this year. The company is founded in 2023 and is owned by hedge fund Zhejiang High-Flyer Asset Management. Its earlier fundraising set a record for first-time financing by a Chinese tech startup. Apart from Tencent and CATL, it notably drew backing of the National Artificial Intelligence Industry Investment Fund, one of the vehicles that spearheads Beijing’s overarching endeavors in the sector.

If you are an executive or board member at a frontier AI company, this story is a warning label wrapped in a funding round. Viral founder posts and unverified transcripts can spill into negotiations, forcing pause-and-restart behavior even during high-demand fundraising windows. The second-order lesson is operational: information hygiene and stakeholder alignment are now part of the fundraising product. DeepSeek’s scramble is less about whether it needs money and more about whether it can control the narrative long enough to close the deal.