DeepSeek founder Liang Wenfeng rejects KPIs, overtime culture as AI lab hits $60 billion valuation

DeepSeek founder Liang Wenfeng said his AI startup, valued at $60 billion, operates without KPIs or overtime culture, a philosophy revealed in a nearly four-hour staff discussion transcribed by Tencent Technology. Founded in July 2023 in Hangzhou, DeepSeek has amassed about 173 million downloads since its public launch in January 2025, with a team of roughly 160 employees, and closed or entered advanced stages of a funding round exceeding $7.4 billion. The company's lean approach challenges assumptions about AI development costs and has impacted crypto markets, with Solana-based meme tokens surging after the R1 model launch.

Via bloomberg.com DeepSeek founder Liang Wenfeng rejects KPIs, overtime culture as AI lab hits $60 billion valuation The Chinese AI startup's unconventional management philosophy stands in stark contrast to Silicon Valley's grind culture, and crypto markets have already felt the ripple effects Liang Wenfeng built a $60 billion AI company in three years. He says he did it without making anyone stay late. The DeepSeek founder revealed his management philosophy during a nearly four-hour discussion with staff, transcribed and obtained by Tencent Technology. His core message was simple: no KPIs, no overtime culture, no performance theater. “We generally don’t work overtime,” Liang said during the session. For a company that has racked up roughly 173 million downloads since its public launch in January 2025, that’s a remarkable statement. How DeepSeek got here DeepSeek was founded in July 2023 in Hangzhou. In under two years, it went from obscurity to directly challenging ChatGPT’s dominance in the consumer AI space. Its valuation reportedly climbed from a roughly $10 billion target in April 2026 to somewhere in the $45 to $52 billion range by mid-2026. The company closed or entered advanced stages of a funding round exceeding $7.4 billion. The team reportedly numbered around 160 employees by 2025. For context, OpenAI employs thousands. DeepSeek achieved competitive AI model performance with a fraction of the headcount, leaning heavily on open-source development and dramatically lower training costs. Initially, Liang financed the entire operation through his hedge fund, High-Flyer. The 2026 funding round marked a strategic shift toward external investment, reportedly including indirect stakes and state-linked investors. That pivot positions DeepSeek as China’s most credible answer to US AI dominance, particularly given the ongoing semiconductor export restrictions that have constrained Chinese tech companies’ access to advanced chips. Why crypto markets care about an AI lab’s HR policies DeepSeek doesn’t have a token. It hasn’t launched any blockchain products. It has no official connection to crypto whatsoever. When DeepSeek’s R1 model launched in January 2025, the shockwave was immediate. Traditional tech stocks with AI exposure took a hit as investors recalculated the competitive landscape. On the other side of the market, Solana-based meme tokens inspired by DeepSeek surged in a speculative frenzy. What this means for investors DeepSeek’s rise challenges the assumption that building competitive AI requires billions in compute spending and massive teams. DeepSeek’s lean operation and emphasis on efficient training suggest the moat around incumbent AI companies may be thinner than their valuations imply. DeepSeek’s open-source approach to AI development aligns philosophically with decentralized infrastructure. Open models lower the barrier to entry for AI-powered decentralized applications, potentially accelerating the convergence of AI and blockchain that projects across the ecosystem have been promising for years. DeepSeek meme tokens are not equity stakes in a $60 billion company. They’re speculative instruments that trade on attention, not revenue. The gap between DeepSeek’s actual value creation and the derivative speculation it inspires in crypto markets is enormous. Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy https://cryptobriefing.com/editorial-policy/ .