DeepSeek Is Taking Outside Money for the First Time — Tencent, CATL, and a State Fund Line Up for a $7.4B Round

The Company That Shook the World Bare-Handed Finally Reached Out

DeepSeek is an unusual company. While most AI startups raised billions to stockpile GPUs, DeepSeek shook the entire global AI map with models like V3 and R1 — without a dollar of outside investment. Released as open source, those models showed you can approach top-tier performance at far lower cost, cracking US Big Tech's "throw money at AI" formula.

Now, in June 2026, that same DeepSeek is nearly closing its first external funding round since founding. The size is about $7.4 billion (RMB 50 billion). With this money in, the valuation is set to land between RMB 350–400 billion, or roughly $52–59 billion. As the moment a "bare-handed David" first accepts giant capital, it carries symbolism beyond a typical funding headline.

The investor list reveals the round's real meaning. Tencent is weighing about RMB 10 billion ($1.4B), battery giant CATL about RMB 5 billion ($700M), and China's national AI industry investment fund is joining. Gaming company NetEase and e-commerce giant JD.com are in late talks too. The investor count is narrowed to fewer than 10, and founder Liang Wenfeng is contributing RMB 20 billion himself — meaning the founder carries the largest share.

The Cast — DeepSeek, Liang Wenfeng, and 'the State'

The first lead is DeepSeek itself. Founded by people out of a hedge fund, it made efficiency its weapon. Without astronomical compute, it produced top-tier models through clever architecture and training techniques, then released them as open source for developers worldwide. That "low-cost, open" strategy made DeepSeek a global star.

Second is founder Liang Wenfeng. The key point is that he's the single largest contributor (RMB 20 billion) in this round. Founders usually get diluted as they take outside capital, but Liang is putting in the most money himself to keep control. The intent reads as "take the money, but I hold the rudder."

The third character is special: the Chinese state. The national AI industry investment fund joining directly, with Chinese champions like Tencent and CATL lining up, means this round isn't a plain market deal but Beijing's de facto endorsement. China treats AI as a strategic national industry, and DeepSeek has become its flagship player. With US chip restrictions on China tightening, the state backing a homegrown AI champion is a clear message.

The Core — The Round by the Numbers and Structure

Item	Detail
Round type	First-ever external funding
Size	~$7.4B (RMB 50B)
Valuation	$52–59B (RMB 350–400B)
Key investors	Tencent (~RMB 10B), CATL (~RMB 5B), national AI fund
Late talks	NetEase, JD.com
Founder contribution	Liang Wenfeng ~RMB 20B
Investor count	Fewer than 10
Reported	June 3–4, 2026; closing within weeks

Three things read off this structure. First, narrowing investors to fewer than 10 signals "we won't take just any money." Second, the founder contributing the largest share to keep control signals "take capital but keep independence." Third, the mix of the state fund with Tencent and CATL means a solid camp backed by both market and state.

One caveat: as of early-June reporting, this is an in-progress deal "closing within weeks." The final investor list, amounts, and valuation could change before close. So treat the figures here not as confirmed but as the late-stage outline of negotiations.

Who Gains What

For DeepSeek, it's dry powder for the next stage of the efficiency strategy. It has done well bare-handed, but training larger models and running more inference infrastructure ultimately needs capital. With latest GPUs hard to get under US chip restrictions, buying domestic compute and infrastructure costs money. The first external round reads as a move to break past "the limits of going it alone."

Investors have different motives. Tencent gains both equity and a rationale to deeply integrate DeepSeek models into its services; a manufacturing giant like CATL secures a foothold to expand into physical-world AI; the state fund gains influence and a strategic asset in a homegrown champion. That such varied motives gathered in one round means DeepSeek is seen as more than a model company — it's viewed as a hub of China's AI ecosystem.

For China's AI ecosystem overall, the round signals a strengthening of the domestic AI stack — part of a bigger plan to build a closed loop that runs on homegrown models, chips, and capital without depending on US models or chips. DeepSeek becomes the software core of that loop.

History — State-Built AI Champions, Hits and Shadows

The success template is tech champions that grew fast on state support. Several countries have poured state capital into strategic industries to grow national champions, some becoming global powers. By adding the thrust of state capital to its distinctive "low-cost, open" edge, DeepSeek could climb to the center of the global open-source AI camp even faster.

But there are shadows. Companies with deep state capital draw worries about being swayed by national strategy over market logic. Overseas customers, especially Western governments and enterprises, may feel security and regulatory burdens adopting "an AI model the Chinese state fund invested in." Indeed, several countries have debated data and security concerns over using DeepSeek models. The source of capital becomes a variable in trust.

The lesson: state capital can be both an accelerator of growth and a shackle on overseas expansion. Whether DeepSeek can enjoy solid state support at home while keeping an "open-source, neutral" image abroad will decide its long-term fate. Those goals subtly conflict.

Rivals' Counter-Play

The US camp (OpenAI, Anthropic, Google) already feels pressure from DeepSeek's low-cost open-source strategy. With state capital behind it, DeepSeek could get more aggressive on both price and the open-source ecosystem. The US camp's counter is widening the gap with top frontier performance, or strengthening positioning as "trustworthy Western models" in enterprise and regulated markets.

Other Chinese AI firms (Alibaba's Qwen, Moonshot's Kimi) have to feel the tension too. The state fund and Tencent betting on DeepSeek can read as Beijing effectively picking its "starting player." Rivals will defend their relevance with differentiation — domain specialization, larger context, multimodality — intensifying the jockeying inside China's AI ecosystem.

Ultimately the big picture is a capital-and-ecosystem contest between the US and Chinese AI camps. DeepSeek's first external round is one step in China completing a loop of "growing homegrown AI with homegrown capital," and the US camp will move to tighten its own camp's cohesion and edge in response.

So What Actually Changes — By Audience

For developers, the core point is that the open-source AI option grows sturdier. With capital secured, DeepSeek gains the means to release stronger models as open source more steadily — a solid alternative to depending only on US models. Just note that in sensitive areas (government, finance, overseas customers), the model's capital source may require security and regulatory review before adoption.

For companies and investors, read the signal that Chinese AI is building a self-sustaining capital loop. Contrary to predictions of isolation under US restrictions, China is rapidly assembling an ecosystem that runs on homegrown capital, chips, and models. Keep in mind the global AI market is splitting from "one market" into "two camps."

For general observers, grab the picture that AI competition has spilled beyond technology into geopolitics. The state fund joining DeepSeek's first external round symbolizes AI becoming a stage for strategic competition between nations, not just companies. Which camp's model you use is increasingly a political choice.

🥄 Three Things You're Probably Wondering

— So what does this mean for me? It matters if you use open-source AI — with capital in hand, DeepSeek is more likely to keep releasing stronger models for free. But if you work with government, finance, or overseas customers, the model's capital source could become a factor in adoption review.

— Why take money now after doing fine bare-handed? Even an efficiency strategy has limits. Training larger models and running inference infrastructure ultimately needs capital. With compute hard to get under US chip restrictions, buying domestic resources costs money. See it as a move to break past "the limits of going it alone."

— Is the state fund joining good or bad? It cuts both ways. At home it's solid thrust; for overseas expansion, the "state-capital-backed model" label can become a security and regulatory burden. It's both accelerator and shackle, so it's too early to call good or bad.

References

• DeepSeek slated to draw $7 billion in maiden fundraising — CNBC
• DeepSeek in talks to raise $7 billion from Tencent, CATL and other investors — TechNode
• China's DeepSeek eyes $7.4 billion in first-ever funding round — American Bazaar
• DeepSeek Eyes $7.4 Billion Funding Round At Up To $59 Billion Valuation — Yahoo Finance
• DeepSeek Pursues $7.4B in Inaugural Funding with Tencent, CATL Leading — MoneyCheck

Numbers and criteria are as of announcement and may change. Investment calls are yours to make!