Joanna Lichter says robotics value is likely to accrue at both ends of the stack
Investor Joanna Lichter outlined a working hypothesis on how value may build in robotics, arguing that the sector resembles a barbell rather than a broad, evenly distributed market. On one side are horizontal robotics foundation models, or RFMs, and infrastructure platforms that can compound through data, compute, and ecosystem effects. On the other are category-leading application companies that capture value by turning intelligence into useful work. In her view, the middle layer looks less attractive, especially where models are replaceable, tools are easy to copy, or deployments lack software leverage. Lichter also drew a contrast with large language models, saying robotics faces distribution costs, scarce physical-world data, real-time compute limits, and safety constraints. Those conditions lead to a gated diffusion pattern instead of immediate mass adoption. She added that deployments can become defensible if they create a learning loop, reduce labor required per robot, and accumulate reusable data and operating knowledge over time. Vertical entry points, she said, can speed up that path.
