DeepSeek Raises $7.4B in First-Ever Funding Round — Tencent and CATL Lead at $59B Valuation

DeepSeek is closing its first-ever outside funding round, lining up approximately $7.4 billion from a group that includes Tencent Holdings, battery giant CATL, and the National Artificial Intelligence Industry Investment Fund of China. Post-money valuation is expected to land between $52 billion and $59 billion — one of the largest private technology financings in China’s history.

The round’s composition is notable. Founder Liang Wenfeng is personally committing roughly 20 billion yuan ($2.75B). Tencent is considering 10 billion yuan ($1.38B). CATL is looking at 5 billion yuan (~$690M). Term sheets are actively being signed, according to Bloomberg and TechNode sources, with the close expected within weeks.

Until now, DeepSeek operated entirely on Liang’s personal capital. No venture funding. No strategic investors. That was a deliberate choice — and it worked. The company behind the V3 and R1 open-source models rose to global prominence in early 2025 by releasing frontier-class AI at a fraction of the training cost claimed by U.S. labs. R1’s January 2025 debut triggered a single-day drop in NVIDIA’s stock that erased approximately $600 billion in market cap as investors recalibrated their assumptions about AI infrastructure spending.

The decision to raise outside capital now signals something has changed in DeepSeek’s calculus. Eighteen months of proving its open-source models work at scale, and the next phase — likely larger training runs, hardware procurement, and product expansion — appears to require capital beyond what one founder can supply alone.

CATL’s presence is the most unusual element of the round. The world’s largest battery manufacturer has no obvious synergy with large language models. Its participation likely reflects either a strategic bet on AI infrastructure’s energy demands, a broader move by Chinese industrial conglomerates to stake positions in strategic technology before competition intensifies, or both.

The fundraise marks a structural shift for the Chinese AI ecosystem. DeepSeek built its reputation by releasing genuinely competitive models without outside capital and without the billion-dollar marketing that U.S. labs deploy at launch. The $7.4 billion raise doesn’t change the models — but it changes the competitive dynamics. A capitalized DeepSeek can operate at a different scale. The open-source releases that disrupted NVIDIA’s market cap in 2025 may have just secured the fuel for their next act.

Sources

• Bloomberg
• TechNode