DeepSeek is reportedly set to raise at least 80 billion yuan — roughly $11.93 billion, and possibly as much as 100 billion yuan ($15 billion) — at a valuation of around $74 billion, with Tencent and battery giant CATL among the largest contributors, according to Bloomberg News reporting. The round would make it one of the largest private AI financings ever, and it lands with a planned IPO on Shanghai's STAR Market in early 2027, a 160,000-chip Huawei-powered data center under construction, and founder Liang Wenfeng's commitment to keep releasing open-weight models. For founders and investors, the DeepSeek funding implications run far deeper than the headline number.
What happened
Per the reports, the raise exceeds 80 billion yuan at an approximate $74 billion valuation, with CATL — the world's biggest EV battery maker — and Tencent anchoring the round. Two facts matter as much as the money. First, DeepSeek is building a data center with more than 160,000 Huawei chips and is developing its own inference chip, a direct push toward compute sovereignty outside the NVIDIA supply chain. Second, its open-weights strategy is intact: founder Liang Wenfeng remains committed to free-weight releases, and the V4.1-Flash model released last month reportedly set new cost/performance benchmarks against models from Anthropic and OpenAI. The planned STAR Market listing via CITIC Securities in early 2027 would give it a public-currency war chest on top of this round.
What the DeepSeek funding implies for founders
The single most important implication: frontier-class open-weight models are going to keep getting cheaper, and their quality floor will keep rising. If you are a founder building on top of models, stop treating the model itself as a durable moat — treat it as a depreciating utility. Build your moat in the things DeepSeek cannot open-source: proprietary workflow data, domain-specific fine-tuning loops, regulatory position, distribution, and the last-mile integration that makes an AI system actually work inside a customer's operation.
There is a second, sharper lesson. DeepSeek is proving that a company can pair open weights with a serious enterprise and infrastructure business at the same time. Founders who default to closed models because "that's where the money is" should re-examine that assumption: the open-weights ecosystem now has a $74 billion champion with a data center the size of a small country's compute fleet. Bet your architecture on open weights becoming a viable — often superior — default, and design your pricing to survive when inference costs fall another 10x.
What it means for investors
Compute sovereignty just became a visible asset class. A battery manufacturer — CATL — anchoring one of the biggest AI rounds in history tells you that energy and compute are now strategic inputs that industrial giants will pay to control, not just costs to be rented from a cloud provider. Investors should be tracking compute-and-energy adjacency deals the way they once tracked cloud-infrastructure plays.
The portfolio question is equally direct: if you hold AI application companies whose pitch is "we have a better model," stress-test them against a world where DeepSeek-quality open weights are free. The survivors will be the ones whose value sits in data flywheels, vertical workflows, or customer distribution. And watch the STAR Market IPO pipeline — a successful early-2027 listing would open a second public-market exit window for AI companies beyond the US, changing the liquidity math for anyone investing in Chinese and cross-border AI startups.
The takeaway
DeepSeek's raise is the strongest evidence yet that open weights and frontier capability are no longer opposites. Founders should build as if top-tier models are a commodity that trends toward free; investors should price compute sovereignty like real estate and diligence AI startups on everything except the model. The money isn't betting on one lab — it's betting that the floor of AI capability keeps rising for everyone.
Angeled.com connects founders with investors.
Read Next
General Translation Raises $27 Million Series A Led by a16z to Build Agent-First Localization Infrastructure
LangChain Raises $300 Million Led by CapitalG at a $3 Billion Valuation to Build the Agent Stack
Mount Thor Raises $45 Million to Run Cloud Fleets of Macs for AI Agents
OpenAI Reportedly Raises $15 Billion at a $1.25 Trillion Valuation
Q-CTRL Raises $290 Million Series C Led by Salesforce Ventures, Reaching Unicorn Status
SE3 Raises $30 Million Series A Led by Accel



