Founder says DeepSeek prioritises AGI over profit, likely to keep top models open-source, Yicai reports

Reuters | July 23, 2026 at 09:31 AM UTC
Bullish 79% Confidence Majority Agreement
Read Original Article

Key Points

  • DeepSeek completed its first funding round at approximately $52 billion valuation, reversing years of rejecting external capital amid pressure from competitors like ByteDance, Moonshot, and Z.ai
  • Liang stated computing power is the biggest constraint and the main gap with U.S. competitors is resource access rather than technical capability
  • The company plans to build large AI computing clusters but has not decided on developing proprietary chips, with Liang preferring to buy chips at reasonable prices rather than manufacture them

AI Summary

DeepSeek Prioritizes AGI Development Over Profit, Maintains Open-Source Commitment

Chinese AI startup DeepSeek is focusing on long-term artificial general intelligence (AGI) development rather than profit maximization, according to founder Liang Wenfeng's remarks to investors during a recent four-hour meeting, as reported by state-owned outlet Yicai on Thursday.

Key Developments:

  • DeepSeek plans to keep its most advanced AI models open-source, with Liang arguing that open-source development and commercial monetization are compatible strategies
  • The company recently completed its first external funding round at a $52 billion valuation, abandoning its years-long policy of refusing outside capital
  • This fundraising shift comes amid intensifying competition from Chinese tech giants like ByteDance and well-funded startups including Moonshot and Z.ai

Technical Challenges and Strategy:

Liang identified computing power as DeepSeek's biggest constraint, stating the primary gap with U.S. competitors is resource access rather than technical expertise. The company plans to build large AI computing clusters but hasn't decided on developing proprietary AI chips. "I hope to be able to buy chips at a reasonable price so that I don't have to make chips myself," Liang stated.

Market Context:

DeepSeek gained global attention in early 2024 after its low-cost V3 and R1 models demonstrated China's ability to compete with U.S. AI leaders despite Washington's export restrictions on advanced chips. The U.S. has limited exports of advanced AI chips to China but has refrained from adding DeepSeek and over 100 other flagged companies to its trade blacklist to avoid escalating tensions with Beijing.

The developments highlight China's AI sector's evolving competitive landscape and the ongoing geopolitical tensions surrounding semiconductor access and AI development.

Model Analysis Breakdown

Model Sentiment Confidence
GPT-5-mini Neutral 80%
Claude 4.5 Haiku Bullish 78%
Gemini 2.5 Flash Bullish 80%
Consensus Bullish 79%