If vision is still waiting for its GPT moment, robotics is a step further back, for a reason that has almost nothing to do with model size. The wall in front of physical AI is evaluation rather than intelligence. You can train a policy overnight, but you cannot test it overnight, because the test happens in the physical world and the physical world runs at one times speed. You cannot speed up reality.
That single fact reorders the whole problem. In digital AI, iteration is cheap: generate, score, repeat, a million times before lunch. In robotics, every iteration costs a real robot doing a real thing in real time, and the scoring is slow, expensive, and hard to standardize. The teams that win will be the ones that industrialize evaluation, not just the ones with the biggest model.
The tyranny of the extra nine
The math is what makes robotics so unforgiving. A task with a hundred sequential steps, where each step succeeds ninety-nine percent of the time, will succeed as a whole only about thirty-seven percent of the time. Ninety-nine percent per step sounds excellent, and it produces a machine that fails roughly two times in three.
Reaching ninety-nine percent success across the whole task requires something closer to four nines, 99.99 percent, at every single step. Each additional nine is a brutal, expensive climb, and each one marks the difference between a demo that goes viral and a product someone will actually deploy. Most robotics you see online lives at one or two nines. The gap between that and a shippable system is years of reliability engineering, not a better prompt.
Data or architecture
There is a real fight about where the unlock comes from, and it splits into two camps:
- Data, argued by Michael Cho of BitRobot. Build the collection pipelines, gather orders of magnitude more real-world interaction, and let scale do for robots what it did for language.
- Architecture, argued just as forcefully by Nils Pihl of Auki Labs. He calls the data-collection gold rush one of the dumbest moments of VC groupthink he has seen, and bets the real bottleneck is design rather than appetite for data.
I find the strongest evidence for the architecture side comes from biology. A mosquito navigates, hunts, and avoids your hand on roughly a tenth of a watt, with no pretraining run and no data center behind it. Whatever it is doing, it is not brute-forced scale. That existence proof suggests a far more efficient approach to physical intelligence is possible, and that pouring more data into the current designs may be climbing the wrong hill.
Where robotics actually is
The honest framing is that robotics is at its GPT-2 moment. It is qualitatively different from a year ago, good enough to be exciting and to convince smart people it is close, but still a full generation short of the breakthrough that turns a research curiosity into infrastructure. GPT-2 was thrilling and basically useless commercially. That is roughly where the robots are.
The last point is the one investors should sit with, because it echoes what is true in vision. The advantage that compounds is access to the real-world interaction data rather than the model itself, and that access turns out to be a commercial and relationship problem rather than a technical one. BitRobot's two-sided market is a bet on exactly this, that whoever can strike the deals to collect real data at scale, in real environments, ends up with the asset nobody can clone. The pipeline into reality is the moat, not the training recipe.