DeepSeek is about to close an $8 billion round that values it at $74 billion, and an IPO is next

DeepSeek went from "cheap AI model that spooked Silicon Valley" to "company about to be worth $74 billion" in under a year, and it's not slowing down.

The Chinese AI lab is closing in on a fresh $8 billion funding round that would value it at roughly $74 billion. That's a serious jump from the $52 billion valuation it landed in June, when it raised over $7 billion in its first ever external round, with founder Liang Wenfeng personally chipping in around $3 billion of that himself.

This second round hasn't had a smooth run. It was paused in late July after some investors got frustrated by leaked comments from Liang, then reopened in early August. The backer list includes battery giant CATL alongside China's state backed National AI Industry Investment Fund, Tencent, and a handful of other big names circling.

Why does a company famous for undercutting everyone on price need $8 billion? Data centres and chips, mostly. Building out its own infrastructure instead of renting compute is expensive, and DeepSeek is reportedly racing to scale up before rivals close the gap it opened with its low cost model earlier this year.

The bigger play is what comes after the cheque clears. DeepSeek is said to be lining up a listing on Shanghai's Star Market, with a filing possibly landing as early as the end of 2026 and a public debut targeted for 2027. If that happens, it puts DeepSeek on a genuine collision course with OpenAI and Anthropic, not just on model performance but on the stock market too.

Remember when DeepSeek's whole pitch was doing more with less? Turns out "less" was a phase. This is DeepSeek playing the long game, and it's betting big money that the world still wants a cheaper alternative to the American AI giants, right up until it becomes one of the giants itself.