Industrial Next wants factory robots to switch tasks in minutes, not weeks

Industrial Next, a San Francisco robotics startup founded in 2021, presented its AI-driven robotic workstations on the Startup Battlefield stage at TechCrunch Disrupt 2026 and used the appearance to announce a partnership with BMW Group Technology Office. The collaboration will evaluate Industrial Next's AI-powered dexterous manipulation for manufacturing use, according to the company. The pitch centers on a persistent complaint about factory automation: robots are typically programmed for one task and one product, so changing a line means reprogramming, physical retooling or swapping equipment, a process the company says can run from weeks to months. Industrial Next claims its stations can be reconfigured in minutes instead, without traditional programming.
At the technical core are self-contained robotic stations powered by what the company describes as proprietary foundation models that tie together vision, task context and motion. The company says new tasks are taught through simulation and demonstration rather than hard-coded scripts, and that its stations can be reconfigured, combined and redeployed across product lines. Founder Lukas Pankau described the approach as giving robots a general understanding of a process, built through a high-fidelity simulation engine close to real-world dynamics, so the machines can perform high-precision work almost without seeing the parts at all.
That framing addresses a known bottleneck in robotics. Unlike language or image models, which can draw on vast internet-scale datasets, physical manipulation lacks comparable corpora; there is plenty of data for making a robot dance, Pankau said, but not for assembling high-precision components. Simulated environments are a common workaround, and the interesting claim here is one of fidelity: simulation is useful for manipulation only when contact dynamics, friction and deformation are modeled closely enough that learned policies transfer to real parts. A likely trade-off is that very high-fidelity simulation is expensive to build and maintain, and its value depends on how well it survives the gap between virtual and physical behavior as tasks and part geometries change.
The founders' backgrounds shape the pitch. Pankau was a lead electrical architect at Tesla before moving to Waymo; Allen Pan led Tesla's autonomous factory efforts and later became chief architect at the autonomous driving company Neolix. The two met at Tesla, in an encounter Pankau recalls as tripping over Pan while Pan slept on a factory catwalk after an all-nighter. They did not initially get along but found complementary skills and a shared fixation on shipping, and kept discussing industrial equipment frustrations after leaving Tesla. Pankau, then at Waymo, described missing Tesla's pace; Pan was in a similar position at the China-based autonomous driving developer Momenta. Roughly three years after leaving, they decided to start the company.
Their stated motivation was the divide between fast-moving vehicle and self-driving software and factory equipment they regarded as decades old. Pankau described being able to run machine learning algorithms to design a car while being unable to program a robot, calling that common among Tesla engineers and saying colleagues were stumped by equipment from the 1980s.
Industrial Next says some of its workstations are already in production making one of the world's most popular smartphones, handling tasks manufacturers had tried and failed to automate for years: threading a delicate flex cable and snapping a tiny connector cap into place. Pankau declined to name the customer. He also described a film-peeling station where the company's robots strip protective film from camera modules and buttons at or above human speed with 100% accuracy, including after being moved to a new phone model and production line they had never seen. Those claims are the company's own and have not been independently verified; the transfer claim is the more consequential one for buyers, since a station that keeps its accuracy on an unfamiliar line is what makes redeployment economics work.
The stated ambition goes beyond phones. Pankau said the company wants to take the technology to more complicated tasks and ultimately to building a full car end-to-end without a human touching it. That is a much larger scope than the concrete deployments described, and it is worth separating the two: precise manipulation tasks such as flex-cable threading and film peeling sit in a narrower band of motions and tolerances than full vehicle assembly, which spans large parts, varied materials and long sequences of operations where a single failure can stop a line.
The BMW partnership is framed as an evaluation, not a production commitment, so its near-term significance is access to a major automaker's manufacturing context rather than a deployed contract. Industrial Next did not disclose pricing, unit economics, deployment counts or timing for broader availability. Pankau also characterized the phone work as the performance the founders wished they had at Tesla, where they said stations were bypassed because they could not be programmed effectively with traditional technologies.
Why it matters: Manufacturers weighing automation now have a concrete alternative to fixed, task-specific robot programming: stations pitched as retoolable in minutes, taught by simulation and demonstration rather than scripts. For robotics and manufacturing engineers, the claim to test is transfer, whether a station keeps accuracy when moved to an unfamiliar product and line, since that determines how much integration work each new task really costs. The BMW evaluation is the signal to watch; if it holds up, the same approach could reach beyond consumer electronics into automotive assembly, though full-car automation without human touch remains the company's long-range goal rather than a demonstrated result.
Based on reporting from the original publisher. Visit the source for full context and later updates.
Publisher excerpt
Industrial Next helps manufacturers automate complex, delicate tasks that have traditionally required human labor.