1/ fact. since 2023, deepseek declined offers from china’s top vcs and tech giants and stayed fully funded by its hedge fund. now it’s opening up to external investors for the first time, raising $300m at a ~$10b valuation.
2/ reasons. the cost structure changed. deepseek’s edge was cheap training, but current models rely on reasoning and agents, which require continuous inference, tool usage, and infrastructure running all the time. this creates ongoing operating costs, not one-time expenses.
3/ constraints. deepseek is building and running infrastructure inside china, which limits flexibility in choosing investors and increases dependence on accessible capital. many western investors are cautious about chinese ai, and cross-border deals face more scrutiny on where models are trained, where data is stored, and where teams operate.
4/ case. the meta acquisition of manus made this visible: manus had moved operations abroad, but regulators still reviewed the deal and restricted founders’ movement, showing that relocating does not remove oversight.
we keep watching
The company that proved you don't need billions to build world-class AI is now asking for money. DeepSeek is raising outside capital for the first time. The target: at least $300 million at a valuation north of $10 billion. Until now, founder Liang Wenfeng funded everything… pic.twitter.com/cwU2pWwVNY
— Chubby♨️ (@kimmonismus) April 20, 2026
Nick Trenkler