The AI funding landscape has been defined by massive Western bets — OpenAI's multi-billion dollar rounds, Anthropic's strategic partnerships — but the most structurally significant capital move of 2026 may be coming from a lab that, until now, wanted nothing to do with outside investors. DeepSeek, the Chinese AI lab that rattled Silicon Valley in early 2025, is finally opening its cap table. And the numbers arriving around that decision are striking.
According to reporting by the Financial Times and Bloomberg, DeepSeek is in active talks to raise its first round of venture capital. In just a few weeks of conversations, its reported potential valuation has climbed from $20 billion to $45 billion. For a lab that has never previously sought outside funding, that trajectory tells its own story about the current appetite for credible AI infrastructure plays outside the US ecosystem.
Why Now, After Resisting for So Long
DeepSeek was founded by Liang Wenfeng, a Chinese hedge fund billionaire who currently controls nearly 90% of the company. The lab's deliberate distance from venture capital was a point of distinction — it signalled confidence in its backing and independence from the performance pressures that come with institutional investors. The Financial Times reports that Liang's hand was forced not by financial pressure but by talent competition. Rivals have been actively poaching DeepSeek researchers, and raising funds allows the company to offer employees equity in the business. That is a retention instrument, not a growth move.
The Round's Architecture
Bloomberg reports that the round is expected to be led by China Integrated Circuit Industry Investment Fund, a state investment vehicle. Also reportedly in talks to participate are Tencent and Alibaba, two of China's dominant cloud operators. The involvement of state capital alongside commercial cloud platforms creates a funding structure that is less about pure return and more about national strategic positioning. This is not a standard Series A.
What DeepSeek Actually Built
The lab rose to prominence in early 2025 after releasing a large language model that trained on a fraction of the compute power and at a fraction of the cost of the leading US models from OpenAI and Anthropic. Since then, it has kept pace with frontier models in areas like reasoning and coding. Its models remain open weight, with versions freely available on Hugging Face under the DeepSeek-AI organisation. That openness has been central to its credibility with technical communities globally.
The Chip Layer Underneath
DeepSeek has been optimised to run on chips made by Huawei Technologies, China's hardware giant. Bloomberg notes that the DeepSeek and Huawei combination is viewed as a strategically powerful pairing for China's goal of building sovereign AI capability that does not depend on US chip supply chains. The funding round, seen through this lens, is less about scaling a startup and more about fortifying an alternative AI stack at the national level. China's investment fund leading the round reinforces that reading directly.
Valuation Speed as a Signal
The jump from a $20 billion to a $45 billion reported valuation over just a few weeks of talks is worth examining carefully. It reflects how rare it is for a lab of DeepSeek's demonstrated capability to enter the market for outside capital for the first time. Investors are not pricing a speculative bet. They are pricing proven model performance, an open-weight distribution strategy that has already built global adoption, and a geopolitical tailwind that makes DeepSeek's independence from US infrastructure more valuable, not less, as chip restrictions tighten.
DeepSeek did not respond to requests for comment at time of publication, according to TechCrunch. The round has not been formally announced, and the figures reported by the Financial Times and Bloomberg remain subject to change as negotiations continue.
If this round closes near the reported $45 billion figure, it would mark one of the most consequential AI funding events not originating from a US lab. The valuation would also land DeepSeek well above many established Western AI companies. For agencies and brands building AI-dependent workflows, the consolidation of Chinese AI capability around a well-funded, state-aligned DeepSeek could accelerate the emergence of a genuinely bifurcated global AI infrastructure — two ecosystems, two chip stacks, and eventually two sets of model defaults shaping creative and marketing technology worldwide.