Moore Threads’ Wang Dong: Embodied AI Firms Face High Chip Costs

Wang Dong's comment reframes the embodied AI cost problem as a hardware dependency issue rather than a pure R&D challenge: firms locked into a single overseas training platform are absorbing a steep domestic cost premium on machines. That squeeze matters most for smaller embodied intelligence startups, which lack the balance sheet to absorb it, and it implicitly argues for domestic compute alternatives like Moore Threads. The open question is whether that cost gap narrows as domestic training hardware matures, or whether it keeps concentrating embodied AI development among better-capitalized players.