In this episode of TechDrive Zurich, I’m taking a drive with deep tech executive and venture partner, Dr. Robert MacKenzie (Ellipsis Ventures, former C-level at ANYbotics, Hocoma, Labster).We dive into what it actually takes to get robotics out of the lab and into the real world. We discuss the transition from flashy demos to achieving "boring reliability", the 99.999% dependability required for industrial scaling. Robert shares insights from his experience scaling ANYbotics into a global player, and breaks down what investors are truly looking for in the next generation of AI and robotics startups.As always on TechDrive, we explore these complex topics while navigating the streets, blending high-level strategy with real-world execution.Three things stuck with me:
1. In deep tech, "if you build it, they will come" is rarely true. Robert's warning is premature commercialization. Chasing revenue with a product that isn't finished. His fix is to stop treating real-world deployments as expensive guesswork ("osmosis," as he calls it, basically put an expensive robot out there and hope you learn something) and turn them into a tight, conscious learning loop that feeds straight back into the product. Build the value first and the adoption follows.
2. The robotics correction is coming. It's a planning signal, not a reason to panic. He's predicting a corrective dip in early 2027 as the hype meets reality and investment cools. His practical advice to Zürich robotics founders was blunt: if you want to raise, raise in 2026. He's not anti-humanoid, but they "do amazing movements with zero understanding of the movements they're doing." The near future is a mixed fleet of specialized robots, not one humanoid doing everything.
3. Founders, especially researchers, shouldn't lose their superpower. Scientists are masters of the learning loop: form a hypothesis, don't know the answer, find a way. But once a spin-off finds one solution that works in the lab, the loop often stops, and they fall in love with it. Keep that loop running in the real world and it's a game-changer. His related, honest note: don't default to CEO/CTO because the title is romantic, but take the role that's best for the company. Sometimes that means letting someone else drive.Subscribe for more in-car conversations with the leaders shaping the future of tech, AI and robotics!
