DeepSeek's AGI-first strategy: founder prioritizes long-term AI research over products and profits

DeepSeek founder Liang Wenfeng held a rare 4-hour investor meeting, with 52 recorded remarks that crystallize the company's singular mission: achieving AGI above all else. Wenfeng stated bluntly that "this is not the time to maximize returns through products" and that building consumer or enterprise products is merely a step toward AGI, not an end goal. He emphasized restraint as a strategy — giving up immediate commercialization in exchange for higher probability of long-term success. DeepSeek open-sources the exact same models it deploys internally, believing that AI's potential scale (eventually 10% of global GDP) makes monopoly futile. The company has no ambition to become the next ByteDance or Tencent; instead, Wenfeng sees team stability as the greatest risk, substantially reduced by the latest funding round.

The meeting also addressed the China-U.S. AI gap, which Wenfeng attributes primarily to resource constraints. DeepSeek trains at current sizes because "these are all the resources we have," adhering to scaling laws that larger models yield better results. He views Anthropic's current lead over OpenAI as temporary, with OpenAI and Google likely trading places in the future. Wenfeng's philosophy: "AGI offers the greatest return. As for everything else, we will do it if we have the capacity, and we will not do it if we do not." This stark prioritization — research over products, open source over capture, stability over growth — positions DeepSeek as a radical bet on long-term AI progress, with no immediate path to commercialization.