In just three years, DeepSeek has gone from an unknown startup to a $60 billion AI powerhouse. The ChatGPT rival has had around 173 million downloads since it launched to the public in January 2025. And its founder Liang Wenfeng says he did it all without a single KPI or late-night shifts from his team.
“We generally don’t work overtime,” Liang revealed in a nearly four-hour discussion session with staff, obtained and transcribed by Tencent Technology.
And the world’s richest AI founder explained that there are two reasons for his laid-back approach: “First, research needs a relaxed environment; if you push too hard, you can’t do research. Second, we’re very focused, which means we do very few things.”
It’s not just the working hours that are relaxed at DeepSeek. There’s apparently no rigid hierarchy handing down orders, no rulebook employees are expected to follow, and no strict written end goal that workers need to adhere to daily. Even how they manage half of their working days is up to them.
“We don’t really have an organization—we’re driven by a vision, organized by a vision,” Liang, who has a $37.9 billion net worth, added. “That vision isn’t even written down. We’ve never written anything.”
