What started in mid-April as a $300m raise at a $10bn valuation, with Alibaba and Tencent talking, is now an FT-reported deal at $45bn led by the China Integrated Circuit Industry Investment Fund. The strategic logic has changed substantially.
There is, in venture-capital pricing, a particular kind of ascent that does not normally happen to a company that has spent its entire existence refusing outside money. On Wednesday, Bloomberg reported, citing a Financial Times scoop, that China’s main state-backed semiconductor investment vehicle is now in talks to lead DeepSeek’s first external funding round at a valuation of approximately $45bn.
The valuation is more than double the figure that was being discussed only two weeks ago. The trajectory, in itself, is striking. The composition of the lead investor is more so.
The China Integrated Circuit Industry Investment Fund, known across the industry as the “Big Fund”, is the central state vehicle through which Beijing has financed China’s semiconductor self-sufficiency push since 2014. I
ts three successive phases have, between them, deployed more than $50bn into Chinese chip-design, fabrication, packaging, and equipment companies. Until Wednesday’s reporting, the fund’s mandate was overwhelmingly focused on the silicon side of the AI stack, fabs, foundries, memory producers, and EDA tooling.
A direct lead investment in a frontier-AI model lab would be, by some distance, the largest extension of the fund’s mandate to date.
It also tells you what Beijing has decided about the strategic relationship between AI capability and chip capability. TNW has tracked the wider US-China chip-export-control dynamic for several years, and the through-line has been consistent: the US has used semiconductor export controls to limit China’s access to leading-edge AI compute, and China has been searching for a response strategy that recovers parity.
The Big Fund leading a DeepSeek round is, in that frame, the recognition that the response strategy now runs through model capability rather than purely through chip capability. If China cannot acquire Nvidia’s leading-edge GPUs at the volume required, it will, on this evidence, finance the model labs that have demonstrated they can produce frontier results without them.
How DeepSeek got here?
DeepSeek’s commercial story is by now familiar. The company was founded in July 2023 by Liang Wenfeng, a 40-year-old computer scientist and co-founder of the quantitative hedge fund High-Flyer Capital Management. Until April 2026, DeepSeek had been funded entirely from High-Flyer’s balance sheet, with no external venture capital and no public communication of revenue. South China Morning Post asked the obvious question last month: why is a company that does not appear to need cash now raising it? The answer, on the available reporting, has two parts.
The first is operational. DeepSeek-R1, the reasoning model launched in January 2025 that triggered the dramatic equity-market reaction in US technology stocks at the time, was trained for a reported $6m, an order of magnitude less than its US peers.
DeepSeek V4, the company’s trillion-parameter flagship, launched on 24 April 2026, the same week that funding negotiations became public. Even with High-Flyer’s balance sheet, training and serving frontier models at the scale DeepSeek now operates is not an indefinitely sustainable internal-funding proposition.
The second is strategic. Yicai Global reported that Liang himself injected fresh personal capital in April, lifting DeepSeek’s registered capital by 50 per cent. His personal shareholding rose from approximately 1 per cent to 34 per cent, with his total combined direct and indirect ownership now around 84 per cent.
The cap-table reorganisation appears to have been deliberately staged ahead of accepting outside capital, both to consolidate founder control and to provide a clean structure for the institutional investors now joining.
From $10bn to $20bn to $45bn in three weeks
The pricing trajectory is worth setting out in order. DeepSeek had quietly opened its first round at a $10bn valuation, targeting roughly $300m. By 22 April, when Bloomberg confirmed Tencent and Alibaba were in negotiations to participate, the valuation had risen above $20bn. By Wednesday’s FT report, with the Big Fund engaged as potential lead, the figure stood at $45bn, more than four times the original opening mark.
Three things have moved that price. The first is investor demand: each new institutional name brought into the round increased the implied valuation that the next investor had to clear. The second is the strategic premium attached to the Big Fund’s involvement specifically, once a state vehicle of this size signals interest, foreign and private investors recalibrate. The third is the wider Chinese AI-funding context, which has tightened markedly in the same window.
We wrote earlier this year on Tencent’s ClawPro launch and the parallel investments by Alibaba, Tencent, and ByteDance in their respective enterprise AI platforms; Alibaba’s AI cloud share is now 35.8 per cent of the Chinese market, Tencent’s 2025 AI spend was ¥18bn with that figure planned to double in 2026, and the broader competitive intensity inside Chinese AI has produced an environment in which a frontier-capability model lab is now treated as a strategic asset rather than a normal venture investment.
A Big Fund-led round would mark a structural shift for DeepSeek. Chinese state-backed investors usually bring more than capital: access to state-aligned customers, regulatory protection, and a signal that the company is now considered strategically important. In DeepSeek’s case, that could mean closer links to central-government AI procurement and state-owned-enterprise demand.
The trade-off is political alignment. A state-backed lead investor may come with expectations around Chinese AI safety rules, content controls, data localisation, and preference for domestic strategic customers. It could also raise questions about whether DeepSeek’s unusually open release strategy can continue unchanged. Neither DeepSeek, the Big Fund, Tencent, nor Alibaba has publicly commented, and the round has not closed.
The wider context is the global AI infrastructure race. As Western governments tighten restrictions around Chinese access to chips and critical technology, Beijing appears to be tightening the financial perimeter around its leading AI companies. The potential investment suggests DeepSeek is moving from self-funded outlier to state-backed national champion.
Three things now matter. First, whether Tencent and Alibaba stay in the round at the reported $45bn valuation. Second, whether DeepSeek keeps releasing model weights and technical reports with the same openness. Third, how the Big Fund’s stake is structured, especially whether it includes strategic governance rights.
On the available reporting, this is the clearest sign yet that Beijing has chosen DeepSeek as one of its central bets in frontier AI. The deal is not closed, but the direction is clear: DeepSeek’s next chapter is likely to be less independent, more strategically aligned, and far more valuable than it was only weeks ago.
Get the TNW newsletter
Get the most important tech news in your inbox each week.