In just three years, DeepSeek has transformed from an unknown startup into a $60 billion AI powerhouse. The ChatGPT rival has recorded approximately 173 million downloads since its public launch in January 2025. Founder Liang Wenfeng attributes this success to an unconventional approach: no key performance indicators (KPIs), no mandatory overtime, and minimal top-down management.

The company, based in Hangzhou and funded by Liang's quantitative hedge fund High-Flyer Capital Management, has drawn global attention for releasing open-weight large language models that developers can download and modify freely—an approach that helps explain the download figures and has made DeepSeek a frequent point of comparison in debates over AI development costs and accessibility.

"We generally don't work overtime," Liang stated during a nearly four-hour internal discussion with staff, the transcript of which was obtained and published by Tencent Technology.

Liang, whose net worth is estimated at $37.9 billion, cited two reasons for the relaxed 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."

The informality extends well beyond working hours. DeepSeek operates without rigid hierarchy, formal rulebooks, or strict daily objectives. Employees reportedly control how they spend half of each workday.

"We don't really have an organization—we're driven by a vision, organized by a vision," Liang said. "That vision isn't even written down. We've never written anything."

Staff members are not assigned monthly targets, and Liang characterized his own leadership as consensus-driven. "No one manages them. No KPIs," he said. "We hope 'formal work' doesn't take more than half of an employee's time. The other half is unallocated. They can do whatever they want."

This structure stands in sharp contrast to the broader Chinese technology sector, where intense work norms remain deeply embedded—and it challenges the prevailing assumption across both Chinese and Western tech circles that winning the AI race requires maximum staffing intensity and capital expenditure.

China's 996 Culture Persists Despite 2021 Ban

China's 996 schedule—working from 9 a.m. to 9 p.m., six days a week—was formally outlawed in 2021. However, the practice has persisted as a point of pride among many technology founders.

E-commerce giant Alibaba, social media titan Tencent, and retailer JD.com have all relied on 996-style schedules as part of their growth narratives. Alibaba founder Jack Ma called the arrangement a "great blessing" in a 2019 essay, framing it as essential to the company's ascent.

The demanding office culture has prompted some young Chinese professionals to leave white-collar careers for lower-paying manual labor. Rising youth unemployment suggests others are withdrawing from the workforce entirely.

Despite these signs of pushback, technology leaders continue to advocate for grueling schedules. Entrepreneurs have taken to LinkedIn to argue that adopting China's 996 model is necessary to remain competitive.

Harry Stebbings, founder of the 20VC fund, intensified the debate last year when he stated that Silicon Valley had "turned up the intensity" and urged European founders to take notice.

"7 days a week is the required velocity to win right now. There is no room for slip-up," Stebbings wrote on LinkedIn. "You aren't competing against a random company in Germany etc but the best in the world."

Martin Mignot, a partner at Index Ventures, echoed the sentiment on the same platform: "Forget 9 to 5, 996 is the new startup standard."

"Back in 2018, Michael Moritz introduced the West to China's '996' work schedule…At the time, the piece was controversial. Now? That same schedule has quietly become the norm across tech," Mignot continued. "And founders are no longer apologizing for it."

Some technology companies are reportedly exploring the addition of 996 clauses in employment contracts, effectively requiring new hires to commit to 72-hour workweeks as the AI race intensifies.

Yet DeepSeek's contrarian approach has propelled it to the forefront of that same competition. While Liang insists the company is run by "a bunch of very ordinary people," his management philosophy is anything but conventional—and whether DeepSeek's model can sustain its momentum as larger, better-capitalized rivals accelerate their own releases remains an open question for the industry.

This story was originally featured on Fortune.com.