Bear Robotics doubles down: 16,000 robots deployed and humanoids rejoin via Kinisi comeback
Key points
Key takeaways from theCUBE on YouTube interview with Bear Robotics Co-founder and Chief Robotics Officer Brennand Pierce (October 2026):
Scale and momentum. Bear has deployed around 16,000 autonomous mobile robots, carries a backlog of about 4,000 units, and says revenue is doubling every year;; the company employs about 380 people, of whom only about 90 are roboticists or software engineers.
Boomerang humanoids. Pierce left Bear to found the humanoid startup Kinisi Robotics; Bear acquired it back, bringing him home as Chief Robotics Officer. The humanoids augment the AMR fleet — pouring drinks, loading robots and owning an end-to-end workflow — and AMRs and humanoids share the same software base and cloud infrastructure.
The model-driven unlock. Foundational models (e.g. Google RT-2, about 3 years ago) let robots learn manipulation from a few hundred examples instead of teams hardcoding positions,and Nvidia Jetson-class compute now runs them in real time onboard.
LLMs in the loop. Feeding an academic paper or blog post into Claude lets Bear replicate its codebase onto its robots within days;; with ChatGPT's Astra,a team closed a complex drawer-opening task in hours using simulation plus real hardware — work that needed about 100 engineers 10 years ago.
Customer first, not tech first. Bear was founded sitting in arestaurant: co-founder John worked in a restaurant and later owned one,and the pair wrote software by dayand tested the new stack on live dinner service each evening, nailing product-market fit in its first year. Pierce says most robotics companies fail not because of technology but because they do not understand what customers value;; his first Munich startup went the wrong way: he built a product and then hunted for customers.
The operational moat. Shipping robots demands hardware, manufacturing, software, sales,and a 24-hour help desk;; Bear built that infrastructure worldwide —and that is why Pierce sold Kinisi back instead of scaling it himself.
Investor realism. Investors split between hype-believers and dot-com-style skeptics asking about ROIand whether robotics becomes a smartphone-like race to the bottom;; Bear's answer is good software plus good hardware,and owning submarkets by understanding the workflow —the robotics equivalent of the iPhone or Nvidia moat.
Where adoption starts. Realistic near-term deployment is back-of-house — restaurant kitchens, hotel laundry, factories,and warehouses —with humanoids serving tables likely five or more years out;; the current race across Europe,the US,and China is technological, not one for supremacy.
Read more: theCUBE on YouTube
