DeepSeek raises $7.4B at $50B valuation in first funding round

Key Takeaways

• DeepSeek raised $7.4 billion in its first external funding round at a $50 billion valuation
• Investors must commit to a five-year lock-up with no voting rights, except for China's state-backed AI fund
• The company's V4 model runs on Huawei chips and undercuts OpenAI's pricing by up to 35x on output

DeepSeek has raised more than 50 billion yuan ($7.4 billion) in its first external funding round, valuing the Chinese AI startup at $50 billion. The deal, reported by The Information, comes with terms that would make most Silicon Valley investors walk away: no voting rights, a five-year lock-up, and money funneled through a limited partnership controlled by founder Liang Wenfeng.

The valuation jumped fivefold in two months. As recently as April, investors were discussing terms around $10 billion. Now DeepSeek sits among the world's most valuable AI companies, though still trailing OpenAI and Anthropic, both approaching the trillion-dollar mark.

Why the unusual deal structure?

DeepSeek's funding mechanism is designed to preserve founder control at almost any cost. Investors put capital into a limited partnership managed by CEO Liang, not directly into the company. They get economic exposure but no board seats, no governance input, no ability to exit for five years.

One exception exists: China's state-backed AI investment fund invested directly into DeepSeek and retains voting rights. That carve-out signals where the company's strategic alignment lies.

Liang himself contributed about 20 billion yuan to the round, according to Reuters. That's roughly $2.9 billion of his own money, reinforcing his majority stake and personal commitment to the company's long-term research agenda.

Who backed the round?

Tencent and CATL lead the roster of outside backers. Tencent brings distribution reach across China's consumer internet; CATL, the world's largest EV battery manufacturer, brings deep pockets and a growing interest in AI for manufacturing and autonomous systems.

DeepSeek was previously funded entirely by High-Flyer, the quantitative hedge fund where Liang made his fortune. This round marks the company's first acceptance of external capital, a shift that suggests DeepSeek's infrastructure and research ambitions now exceed what internal funding can sustain.

What DeepSeek is building

The company made global headlines in early 2025 with its V3 and R1 models, which matched or exceeded Western frontier models on key benchmarks at a fraction of the training cost. In April 2026, DeepSeek released V4, the largest open-weights model to date. It runs entirely on Huawei chips, a notable detail given U.S. export restrictions on advanced semiconductors to China.

Liang has told investors he prioritizes foundational AI research and AGI development over short-term profits. The company plans to keep releasing open-source models, a strategy that builds developer mindshare while pressuring closed competitors on price.

That pricing pressure is already severe. DeepSeek made its 75 percent discount on V4 Pro permanent, making the model roughly 11 times cheaper on input and 35 times cheaper on output than OpenAI's GPT-5.5. For developers and enterprises running inference at scale, the cost difference is hard to ignore.

What this means for the AI race

DeepSeek's $50 billion valuation is modest next to OpenAI's near-trillion figure, but the gap in valuations masks a narrower gap in capabilities. The company has demonstrated that frontier-class models can be built outside the U.S. semiconductor ecosystem, using Huawei's domestically produced chips.

The governance structure matters here. Western AI labs face increasing pressure from investors, regulators, and safety researchers to slow down, open up, or change course. DeepSeek's structure insulates Liang from all of that. He controls the partnership. He sets the research agenda. Outside investors are along for the ride, with no mechanism to steer.

Community reactions on HackerNews and Reddit have focused heavily on the governance model. Some view it as a red flag for accountability; others see it as the kind of founder control that lets research organizations take long-term bets without quarterly pressure. Both readings are probably correct.

Frequently Asked Questions

How much did DeepSeek raise in its first funding round?

DeepSeek raised more than 50 billion yuan, approximately $7.4 billion, in its first external funding round.

What is DeepSeek's current valuation?

The company is now valued at $50 billion, up from $10 billion in discussions just two months earlier.

Why do DeepSeek investors have no voting rights?

Investors contributed to a limited partnership managed by CEO Liang Wenfeng, not directly to DeepSeek. This structure preserves founder control while limiting investor governance.

Who are DeepSeek's major investors?

Tencent and battery maker CATL are among the largest outside backers. China's state-backed AI investment fund also invested directly and retains voting rights.

How does DeepSeek pricing compare to OpenAI?

DeepSeek's V4 Pro is roughly 11 times cheaper on input and 35 times cheaper on output than OpenAI's GPT-5.5.

Source: The Decoder / Jonathan Kemper