There’s a quote from Liang Wenfeng — the founder and CEO of DeepSeek — that got passed around a lot when his lab became the most talked-about AI company on earth in early 2025. He’d said it back in 2023, and it captured something distinctive about how DeepSeek operated: “Money has never been the problem for us. Bans on shipments of advanced chips are the problem.”

That was the philosophy. No outside investors. No venture capital. No state money. The company was funded entirely through profits from High-Flyer, Liang’s quant hedge fund, and Liang liked it that way. He’d told associates he feared outside investors would interfere with DeepSeek’s decisions. He was particularly wary of government-linked funds. He owned 84% of the company and had no plans to change that.

On June 3, 2026, Reuters and Bloomberg reported that DeepSeek is set to raise approximately 50 billion yuan — $7.4 billion — in its first-ever external funding round, at a post-money valuation of $52–59 billion. Term sheets are already being signed. The round is expected to close within weeks. Something changed. The question is what.

Why Is DeepSeek Raising Money Now?

Liang’s original position was coherent. High-Flyer was profitable enough to fund DeepSeek’s research. The constraint wasn’t capital — it was access to Nvidia’s most advanced chips, blocked by US export controls. You can’t buy your way around a chip ban, so why dilute ownership for money you don’t need?

Three things appear to have changed that calculus.

1. Talent poaching became a real threat

According to the Financial Times, the primary trigger for the fundraise was competitors aggressively recruiting DeepSeek’s researchers. Without equity in the company, DeepSeek’s engineers had no financial stake to hold them in place against offers from better-funded rivals. The funding round allows Liang to offer employees shares — the same retention tool that every well-funded startup uses. It’s a mundane, practical reason that has nothing to do with ideology. The company that built an AI revolution on pure research instinct now needs stock options like everyone else.

2. The compute gap is widening

DeepSeek’s efficiency breakthroughs — doing more with less hardware — were genuinely remarkable. But they were also a response to necessity. US export controls mean DeepSeek cannot buy Nvidia H100s or H200s. It has been optimised to run on Huawei’s Ascend chips and older Nvidia hardware. That constraint drives innovation up to a point, but the frontier models from OpenAI, Anthropic, and Google are running on clusters of hundreds of thousands of the most advanced chips available. The gap between constrained-efficient and unconstrained-powerful is growing. More capital means more domestic compute — Huawei Ascend clusters at scale, and potentially China’s own next-generation silicon.

3. Agentic AI requires infrastructure money

DeepSeek is now expanding into agentic AI — software that can perform multi-step tasks autonomously without human intervention. This is the next frontier for every major AI lab. Building production-grade agentic systems requires inference infrastructure, server capacity, and API reliability at a scale that High-Flyer’s profits alone may not sustain indefinitely. The $7.4 billion buys runway into this next phase.

Who Is Investing — and Why Each Investor’s Motive Matters

The investor lineup tells its own story. Fewer than ten investors in total — a deliberate choice to limit outside influence. But each has a distinct strategic reason for being in the round.

The most significant name on that list isn’t Tencent — it’s the National Artificial Intelligence Industry Investment Fund, backed by the China Integrated Circuit Industry Investment Fund (known as “the Big Fund”) — the same state vehicle that funds China’s largest chip companies including SMIC. Liang previously went out of his way to avoid government-linked investors, citing concerns about the appearance of Beijing influence. The fact that China’s state AI fund is now participating is either a pragmatic concession or a signal that the line between private and state AI development in China is blurring further. Probably both. Notably, Alibaba — which was in early discussions — ultimately did not invest, suggesting not every Chinese tech giant saw the same strategic value in the deal.

What DeepSeek Told Investors

According to Bloomberg, DeepSeek’s senior management told potential investors that the company will focus on groundbreaking AI research over short-term commercialisation. Liang pledged in at least one investor meeting to continue developing open-source AI models while pursuing the broader goal of achieving artificial general intelligence. The Hugging Face CEO, Clement Delangue, posted publicly that the open-source commitment remains intact — the MIT licence on DeepSeek’s models is not changing.

That’s a remarkable pitch. Most companies raising $7.4 billion tell investors about their monetisation roadmap. DeepSeek is essentially telling investors: we’re going to keep giving the models away for free, we’re going to focus on the hardest research problems, and we’re going to try to build AGI. Trust us.

The fact that investors are signing term sheets suggests they believe that a free, open-source, frontier AI model from China is itself a strategic asset worth billions — regardless of whether it generates direct revenue. In a world where AI infrastructure access is a geopolitical lever, owning a piece of the leading open-weight Chinese model isn’t a bad position to be in.

The Geopolitical Layer

DeepSeek accepting money from the National AI Industry Investment Fund will not go unnoticed in Washington. The company has already been banned in Italy, South Korea, Australia, Taiwan, and by several US government agencies. The rationale has always been data privacy under Chinese law — concerns that Beijing can compel access to data processed by Chinese AI systems.

State fund participation strengthens the argument that DeepSeek is part of China’s strategic AI infrastructure, not just an independent research lab. That’s unlikely to open new markets in Europe or the US. It will almost certainly accelerate further government restrictions. Ironically, the funding that DeepSeek needs to compete at the frontier may also cement the geopolitical ceiling above which it cannot grow in Western markets.

My Take

The most interesting thing about this funding round isn’t the money. It’s what the money represents: the end of DeepSeek’s exceptionalism. For two years, the company operated on a genuinely different set of principles — no investors, no monetisation pressure, pure research, open source. That made it unusual not just in China but globally. It was the AI lab that didn’t play the funding game and still produced models that made OpenAI nervous.

Now it’s playing the funding game. The talent retention argument is understandable. The agentic AI infrastructure argument is understandable. But the moment you take $7.4 billion from Tencent, CATL, and a state AI fund, you are no longer just a research lab. You are a strategic asset in China’s technology ecosystem. The investors know that. Beijing knows that. The question is whether Liang Wenfeng can hold his original philosophy — AGI-first, open-source, research over monetisation — while managing those stakeholder dynamics.

His own $2.94 billion personal contribution suggests he’s trying. Contributing 40% of the round from his own pocket is how you tell your new investors who’s still in charge. Whether it’s enough to keep DeepSeek’s soul intact alongside its balance sheet is a question only the next few years will answer.

Mr Wangdoo

Founder and Editor-in-Chief of Wangdoo.com. Independent tech journalist covering AI, EVs, gadgets, and emerging tech.