DeepSeek once challenged ChatGPT on a tiny budget. Now investors are lining up.

DeepSeek used to be the quiet kid in the AI classroom.

Before it became a global name, the Hangzhou-based startup was mostly known inside China’s tech circle, backed by founder Liang Wenfeng and his quant fund High-Flyer. No giant marketing machine, no massive Silicon Valley-style fundraising rounds, no “please look at us” energy.

Then DeepSeek dropped its open-source reasoning model and suddenly everyone was looking. The reason was simple: it showed that a Chinese AI team could build a model competitive with top global players while reportedly spending far less on training than many people expected. Reuters later reported that DeepSeek said its R1 model cost just $294,000 to train, a number that restarted the whole debate around whether AI progress really needs endless billions and endless chips.

Now comes the plot twist: DeepSeek is reportedly close to its first external fundraising round, with a possible valuation of up to $50 billion. Reuters reported that the round could raise $3 billion to $4 billion, with China’s state-backed Big Fund involved and Tencent also in talks.

So the big question is not just “can DeepSeek survive?” anymore. It is: if DeepSeek already scared the global AI industry with limited funding, what happens when it finally gets serious money?

Can it become the world’s No.1 AI company? Too early to say. OpenAI, Google, Anthropic, Meta and others are still moving insanely fast. But DeepSeek has already proved one thing: sometimes the most dangerous player is not the one with the biggest budget — it is the one that learned how to fight without one. 🧠🔥