DeepSeek is in talks for its first ever venture round, and the price tag has roughly doubled in a matter of weeks. TechCrunch AI reports that the Chinese AI lab’s potential valuation has climbed from $20 billion to $45 billion, citing the Financial Times and Bloomberg. The round is said to be led by China Integrated Circuit Industry Investment Fund, the country’s state investment vehicle, with Tencent and Alibaba reportedly in talks to participate.
Why DeepSeek finally took the money
Founded by hedge fund billionaire Liang Wenfeng, who still controls nearly 90% of the company, DeepSeek had previously avoided outside investors entirely. What changed? Talent. According to TechCrunch AI, competitors started poaching DeepSeek researchers, and Liang opted to raise so the lab could hand out equity to employees. That’s the practical trigger here. The funding isn’t framed as a compute play or a GPU shopping spree. It’s about retention.
The geopolitical layer
DeepSeek isn’t just another lab. The company shot to global attention in early 2025 after releasing a large language model trained on a fraction of the compute and at a fraction of the cost of OpenAI’s or Anthropic’s frontier systems. It’s stayed competitive on reasoning and coding while keeping its weights open on Hugging Face, which made it the default reference point for anyone arguing that frontier AI doesn’t have to be a closed, U.S.-only club.
The state-backed lead investor signals something bigger. China is funneling capital into homegrown AI to route around U.S. chip restrictions, and DeepSeek has been optimized to run on Huawei silicon. That pairing of a frontier-class model running on domestic chips is exactly the stack Beijing wants to scale.
Why this matters
A few things stand out:
- Open weight stays well funded. The biggest open-weight player in the world is about to get a war chest. Good news for anyone building on top of DeepSeek models.
- The cost narrative gets reinforced. DeepSeek’s pitch was always “frontier results without frontier compute spend.” A $45B valuation says investors believe that thesis holds.
- Talent is the new bottleneck. When even DeepSeek, run by a billionaire founder with no need for outside cash, has to raise just to keep researchers around, that tells you where the squeeze is across the industry.
- The U.S./China stack split widens. Huawei chips plus DeepSeek models plus Tencent and Alibaba distribution is a complete, self-contained AI stack. That ecosystem is now well capitalized.
What to watch
DeepSeek didn’t respond to TechCrunch AI’s request for comment, and terms haven’t closed yet. Two threads worth tracking. First, whether Tencent and Alibaba actually come in, which would put the country’s two biggest cloud players on the same cap table as a state fund. Second, whether the next DeepSeek release leans harder into Huawei optimization as part of the deal narrative.
If the round closes near $45 billion, expect U.S. labs to face renewed pressure on both pricing and openness. The compute-efficiency story DeepSeek pioneered isn’t going away. It’s getting funded. Full details at the original source.