Why Did DeepSeek Suddenly Suspend Its Second Funding Round?

DeepSeek has put its second fundraising round on hold. The Chinese AI startup told prospective investors directly — the ones already approached about the deal — that the process has been suspended.

Publicly, there’s no timeline for when it might resume.

Also Read: Korea’s 200MW AI Factory Hides a Stunning $10 Billion Financing Play

According to Reuters, citing Bloomberg News, the timing is odd.

DeepSeek had been riding extraordinary global momentum. Its January release sent Nvidia‘s stock down 17% in a single session and forced a broad reassessment of AI compute economics around the world.

DeepSeek Funding Pause Upends China’s AI Capital Cycle

DeepSeek is the Hangzhou-based research lab behind the R1 and V3 model families.

Unlike most frontier AI labs, it operates without disclosed venture backing and has historically been funded by its parent, the quantitative hedge fund High-Flyer Capital Management.

The second fundraising round would have been the first time DeepSeek sought capital from external investors beyond its founding structure. The DeepSeek funding pause suggests either that internal disagreements over valuation persist, that prospective investors have raised concerns about the company’s legal or geopolitical exposure, or that DeepSeek’s leadership has pulled back from external ownership altogether.

No valuation figure for the paused round has been reported.

Earlier this year, analysts estimated DeepSeek could command a valuation above $20 billion given benchmark performance comparable to OpenAI’s best models at a fraction of the training cost.

The January Shock That Made DeepSeek Worth Watching

DeepSeek’s R1 model, released publicly on January 20 this year, generated immediate panic in Western technology markets. The model matched or exceeded OpenAI’s o1 on several reasoning benchmarks while DeepSeek claimed a training cost of roughly $6 million, a figure that called into question the assumption that frontier AI requires billions in compute spending.

Nvidia (NVDA) lost approximately $600 billion in market capitalization in the trading session that followed, the largest single-day loss for any company in stock market history.

The event forced a public rethinking of AI capex projections at Microsoft, Google, and Meta.

DeepSeek’s V3 model, released shortly after, extended the company’s lead in open-weight Chinese models. V3 scored competitively on coding and mathematics benchmarks and was made freely available for download, an unusual move for a lab of its caliber.

The combination of benchmark performance, low reported costs, and open availability gave DeepSeek a narrative profile that no other non-US lab had achieved since the current AI cycle began.

Why External Investors May Have Hesitated

The DeepSeek funding pause reflects complications specific to the company’s situation.

DeepSeek operates under Chinese jurisdiction, meaning investor protections would depend on Chinese contract law rather than US or Cayman Islands structures that global technology investors are accustomed to.

US investors face additional constraints. The Treasury Department’s outbound investment rules, finalized in late 2024, restrict Americans from taking stakes in Chinese AI companies that develop frontier models above defined compute thresholds.

DeepSeek almost certainly clears those thresholds.

European and Middle Eastern sovereign funds have shown appetite for Chinese AI exposure, but DeepSeek’s opacity around its actual compute infrastructure and data sourcing practices has been a consistent concern. The company has never published a technical report disclosing its full training hardware stack, and US export controls on advanced Nvidia chips make the provenance of its compute a sensitive question.

There is also a regulatory risk specific to the model weights themselves.

Several US lawmakers have introduced bills that would restrict distribution or use of Chinese-origin AI models on national security grounds. A fund taking an equity stake in DeepSeek could find itself in a politically exposed position if those bills advance.

What a Sustained DeepSeek Funding Pause Would Mean for the AI Funding Landscape

The DeepSeek funding pause matters beyond the company itself.

Chinese AI investment has accelerated sharply in 2026, with domestic investors funding a wave of model companies attempting to follow DeepSeek’s cost-efficiency approach. A freeze at the flagship name sends a cautionary signal about how far that capital cycle can extend.

For Western AI labs, the DeepSeek funding pause temporarily removes one source of competitive pressure.

A capitalized DeepSeek with external backers would have had stronger incentives to commercialize aggressively and challenge OpenAI and Anthropic in enterprise markets. A DeepSeek that remains essentially a research operation inside a hedge fund is a different kind of competitor.

The decision also raises questions about whether China’s most technically capable AI labs can attract the scale of capital flowing to US counterparts. OpenAI raised $40 billion in a single round earlier this year.

The DeepSeek funding pause suggests the structural advantages of the US funding ecosystem remain intact even as the model performance gap narrows.

How DeepSeek Got Here: High-Flyer’s Unlikely Lab

High-Flyer Capital Management founded DeepSeek as an internal research initiative in 2023, allocating significant GPU resources from a cluster the fund had assembled for its own quantitative trading work. The lab was never designed as a venture-backed startup.

That structure gave DeepSeek unusual freedom.

Without quarterly reporting obligations or investor pressure to ship products, the team could pursue research directions that a commercial lab might deprioritize. The result was R1, a model that combined chain-of-thought reasoning with reinforcement learning in ways that surprised even researchers at competing labs.

The fundraising round, when first reported in spring this year, was described as a move to accelerate hiring and expand inference infrastructure to meet surging demand for the R1 API.

Those pressures presumably remain. Whether the DeepSeek funding pause is resolved through internal High-Flyer capital or through a resumed external process is now an open question.

Read Next: Homeowners Could Soon Host The Computing Power Behind AI