DeepSeek Seeks $20B+ Valuation in First-Ever Funding Round

DeepSeek, once the scrappy lab that rattled Silicon Valley with cheap, powerful models, is now raising outside money for the first time — and the valuation has already doubled from an initial ~$10B target to $20B+ in days, driven by intense investor demand.

But the backstory is messier than the headline. V4 is delayed, key researchers are being poached by Tencent, Alibaba, and ByteDance, and the team has burned months porting models to Huawei and Cambricon chips — while still quietly training on smuggled Nvidia Blackwell GPUs. Compute costs are skyrocketing.

The irony: DeepSeek is now valued below MiniMax ($30B+), a lab it once outshone. And its two biggest investors-in-waiting are the same giants raiding its talent. The lab that declared money was never the problem is now raising from the very companies making it a problem.

Why it matters: DeepSeek's funding round signals that even China's most independent AI lab can't escape the gravity of Big Tech — and that the era of scrappy open-weight dominance may be giving way to a familiar consolidation playbook.