At the APEC summit in Shenzhen (November 17–18, themed " " — "Promoting Prosperity Together, Winning the Future Together"), Xiaomi founder Lei Jun laid out an uncomfortable thesis: over the past four years, large models have made huge leaps, but there are still few scenarios where AI actually makes money, and only a handful of companies have a closed commercial loop. His real bet, he said, is embodied intelligence — robots with a physical body, whose market "imagination capacity," in his words, will exceed that of the auto industry.
The debate in the thread isn't really about robots — it's about the first half of that statement. The admission that the industry still can't convert technological progress into profit cuts against the entire rhetoric of recent years — aimed at investors, developers, and everyone selling "AI implementation" as a finished product.
The second half — prioritizing the physical world over the digital — isn't consensus either: some participants see pragmatism in it, others see an attempt to shift the hype cycle to the next thing.
What data underpins the assessment of AI companies' profitability, and how it relates to Xiaomi's own investments in robotics, is unknown; the source material doesn't include the full text of the speech or reactions from summit participants.
1. If AI companies barely make money, is that a failure of business models or just the early stage of a platform technology?
2. Are the robot and auto markets comparable in demand structure — or does the comparison set investors up with the wrong expectations?
3. What has to change in the product for AI to stop being sold as a pilot and become a purchase with a clear price and payback?
Source: zhihu.com
Source: www.zhihu.com · post in Telegram