Opportunity assessment: high. Factory automation has a flexibility problem, and a San Francisco startup says it has fixed it. Industrial Next, founded by two former Tesla engineers, presented today on the Startup Battlefield stage at TechCrunch Disrupt 2026. According to TechCrunch AI, the company builds self-contained AI robotic stations that it says can retool a production line in minutes instead of weeks, with no traditional programming.
It also announced a partnership with BMW Group Technology Office. BMW will evaluate Industrial Next’s AI-powered dexterous manipulation for manufacturing.
🎯 The Problem
Most factory robots do one job. They’re programmed for one task and one product. If the product changes, someone has to reprogram the robots, retool the line or replace equipment, and that can take weeks or months.
That rigidity costs more than time. Pankau says even a small change in a process or product can make a traditional robot fail or cause expensive downtime. Manufacturers end up simplifying their products just to keep them repeatable. Delicate, complex jobs stay manual.
⚙️ What Industrial Next Built
TechCrunch AI lists the main parts of the system:
- Proprietary foundation models. These connect vision, task context and motion in one system.
- Simulation-first training. Robots practice in a high-fidelity virtual world that closely copies real physics. Pankau says this lets them do high-precision tasks “almost without seeing the parts at all.”
- Modular stations. Units can be reconfigured, combined and moved between product lines.
- Teaching instead of coding. New tasks come from simulation and demonstration, not hard-coded scripts.
📱 Already in the Field
Pankau says some of the company’s workstations already help build one of the world’s most popular smartphones. He didn’t name the customer. The stations handle jobs manufacturers had tried and failed to automate for years:
- Threading a delicate flex cable
- Snapping a tiny connector cap into place
- Peeling protective film off camera modules and buttons
The company says the film-peeling station works at or above human speed with 100% accuracy. It reportedly kept that up even after being moved to a phone model and production line it had never seen.
🧠 Why This Matters
What stands out here is the data strategy. Large language models learned from the internet. Robots have nothing that big to learn from. “We can make robots do all these cool things like dance around, because there’s data for that, but we can’t make them assemble high-precision things,” Pankau said.
Industrial Next’s answer is to generate the data in simulation. Physical AI players such as Nvidia are betting on the same idea. If high-fidelity simulation really carries over to sub-millimeter assembly work, the biggest bottleneck in robotics gets much smaller.
The team’s background helps too. Lukas Pankau was a lead electrical architect at Tesla before he moved to Waymo. Allen Pan led Tesla’s autonomous factory efforts and later became chief architect at Neolix. They met when Pankau literally tripped over Pan, who was asleep on a factory catwalk after an all-nighter.
“I could run machine learning algorithms to design a car for me. But I couldn’t figure out how to program a robot,” Pankau said. He described engineers “stumped at the factory looking at this equipment from the 1980s.”
⚠️ Caveats
- Unverified claims. The 100% accuracy figure and the smartphone deployment come from the company. The customer is undisclosed.
- BMW is an evaluation, not a rollout. It’s a pilot-style partnership, not a production contract.
- No pricing or availability details. The article doesn’t say how manufacturers can buy or deploy the stations.
- The gap between simulation and reality. Plenty of robotics startups have hit trouble when sim-trained behavior met the messiness of real factory floors.
🔭 What Comes Next
The founders are aiming high. “We ultimately want to take this to building a full car end-to-end without any human touching ever,” Pankau said.
That’s a long way from peeling film off phone parts. Still, the BMW evaluation gives Industrial Next a real automotive testbed. If the stations hold up there, flexible automation could move from pitch decks to assembly lines. You can find more details in TechCrunch AI’s original coverage.