DeepSeek Founder Liang Wenfeng Says Open-Source Strategy Supports Six-Fold Profit

In a recent internal exchange with investors, Liang Wenfeng, the founder of the Hangzhou-based AI lab DeepSeek, provided a rare look into the financial philosophy driving one of the world’s most watched artificial intelligence companies. Addressing concerns that open-sourcing frontier models might cannibalize revenue, Liang argued that the company’s current business model remains robust under an open-source framework, provided the goal is sustainable growth rather than extreme, short-term profit extraction.

The Financial Math of Open-Source AI

Liang’s strategy relies on a specific financial target: aiming for a six-fold return on investment, which he calculates aligns with a ten-month payback period.

We will definitely open source, and the strongest models will also be open sourced—because I don’t see any necessary benefit to closed source. Liang Wenfeng, DeepSeek Founder, via Mydrivers

However, Liang acknowledges that open-sourcing creates a hard ceiling for those seeking “100-fold” profits. He explains that if a company attempts to charge exorbitant prices, the existence of an open-source alternative allows third parties to deploy the technology independently at twenty times the cost, effectively forcing prices down.

Market Valuation and Personal Wealth

The company’s approach to growth has yielded significant returns for its founder. In June, DeepSeek closed a $7.4 billion funding round that pushed its valuation to approximately $50 billion. This financial surge has had a direct impact on Liang’s personal fortune.

Liang’s position as the wealthiest founder in the AI sector is distinct due to his high level of equity.

The “Sweet Point” of Mid-Size Operations

Beyond the balance sheets, Liang argues that DeepSeek occupies a “sweet point” in the industry. He suggests that small startups lack the raw strength to tackle fundamental research, while massive corporations often suffer from bureaucratic inertia that hampers rapid innovation. By maintaining a mid-sized, highly focused team, Liang claims the company avoids the need for excessive overtime or the “hard” mode of development that plagues larger competitors.

Reuters reported that while Italy and Czechia have placed restrictions on the remote use of DeepSeek’s models, they continue to permit locally hosted deployments. This indicates that while the geopolitical tension surrounding Chinese AI remains high, the utility of the company’s open-source output continues to influence the global infrastructure strategy of governments seeking alternatives to proprietary U.S. systems.

As the company pursues artificial general intelligence, the core question remains whether this six-fold profit model can withstand the pressures of a $50 billion valuation and the scrutiny of international markets. For now, Liang continues to prioritize technical momentum, betting that the ability to release high-performance, open-source models will prove more valuable in the long run than the short-term gains of a closed ecosystem.