This transcript is from a 20-minute keynote delivered at the Munich Urban Colab on July 14th, 2026, to an audience of leading robotics and AI entrepreneurs, founders, and scientists.

—

Hi everyone, dear robotics community,

Thank you for the warm welcome, and thank you, Nicole, for inviting me.

It is a privilege and a pleasure to speak to you today. I am excited and grateful to see so many of you here.

When Munich Urban Colab started its operations on June 28th, 2021—five years ago—we believed that robotics and autonomous systems would change the world.

But little did we know.

Five hackathons and many pizzas later, this lab has supported 78 teams, which have raised more than €64 million in pre-seed and seed capital. What a success! And this gives me great hope for the economic future of Bavaria.

Outside of this lab, the robotics and AI landscape is developing boldly.

Yesterday, Helsing, Europe’s leading defense AI company from Munich, announced that it raised $1.8 billion in a Series E round, valuing the company at $18 billion.

Earlier this month, Proxima Fusion raised €411 million to build Europe’s Commercial Fusion Champion and power the robotics and AI economy.

And last month, Neura Robotics—the pioneer in humanoid robotics near Stuttgart—announced a record Series C funding round of $1.4 billion.

These are the biggest deep-tech funding rounds Germany has ever seen. They prove one thing: Germany is making bold bets on energy, robotics, and AI.

Going forward, I’d like to focus my speech on humanoid robotics.

Let’s look at the US first. Firms such as Figure AI, Boston Dynamics, Tesla, and Apptronik have attracted more than $100 billion for their humanoid robotics programs. American startups are able to deploy vast resources and scale humanoids quickly.

My second example is China. China leads manufacturing at scale, thanks to a mature supply chain and low production costs. UBTECH Robotics, Unitree Robotics, and AgiBot produced more than 10,000 units in 2025. As a result, Chinese firms accounted for an 80% market share of all shipped humanoids in 2025.

The quality and practical use cases of these humanoid systems still need to be evaluated. But we can assume Chinese firms are working tirelessly on both as we speak.

And yes, we all know the funny videos on social media of humanoid robots falling flat during a boxing match or collapsing while playing soccer. But we also know the capabilities of the most innovative systems—the ones that can dance, jump, run a marathon, or beat us at Go.

With technological advancements, large-scale global investments, and growing manufacturing capacity, it is no longer a question of if humanoid robotics will enter our lives—it is a question of WHEN.

This raises a couple of key questions.

Yes, robotics will be a huge help to us in our households, entertaining kids and the elderly (hi, Robody!), or working in our industries. They are going to be an extremely powerful new tool for humanity, allowing us to delegate undesirable tasks.

But what happens when humans and humanoids face an alignment problem? When our interests and values come into conflict?

What happens to society when a few extremely powerful and wealthy firms—and markets tend to consolidate into oligopolies or monopolies—own large fleets of humanoid robotic systems and the associated datasets?

Who holds them accountable?

And by accountable, I mean in terms of paying taxes, which we need to fund governmental services like education, infrastructure, and healthcare. Humanoids, or any artificial system for that matter, do not pay taxes.

Secondly, who holds firms accountable when their products cause accidents?

I’d like to share an example involving a Tesla Model S on Autopilot (not a humanoid, but relevant).

On the night of April 25th, 2019, George McGee was driving his Tesla Model S on Autopilot along Card Sound Road in Florida. When he reached for his dropped phone and looked up too late, he crashed into a Chevy Tahoe parked on the side of the road.

Dillon Angulo, 27, who had pulled off the unlit road to look at the stars with his girlfriend, Naibel Benavides, 22, was left with a shattered pelvis, a broken jaw, and a bleeding brain. Naibel Benavides died at the scene; her body was found over 115 feet from the truck.

A system on Autopilot can make mistakes, just as any human can. And Tesla’s own records confirmed that every vehicle with Autopilot automatically transmits crash data to the company’s servers after a collision.

BUT in this case, the software on the Tesla Model S deleted the video material. The police received only corrupted, unreadable data.

During the investigation, Tesla later stated it had never received the full crash dataset and could not produce any insights into the Autopilot system’s activity.

However, investigators managed to restore the data with the help of experts directly from the Autopilot computer at a Florida Highway Patrol evidence facility in October 2024—five years after the crash.

What they found contradicted Tesla’s account entirely.

The data showed that the system had actively identified the stop sign, the parked vehicle, and even caught a glimpse of the two humans in its path. But it issued no forward collision warning. Automatic emergency braking was never triggered. In the final seconds, Tesla’s Autopilot flashed a „take over immediately“ prompt and surrendered control to a driver who had no time to react.

The system had seen everything and done nothing.

In the final court case, Tesla was ordered to pay $200 million in punitive damages. It was the largest verdict ever handed down against Tesla and the first time a jury found the company liable in a wrongful death case tied directly to Autopilot’s operation.

I mention all of this because when a publicly traded tech company like Tesla does not take responsibility for its powerful and autonomously operating products when damages or fatalities occur, then the justice system as we know it is in jeopardy.

(PAUSE)

Now, I’d like to talk about what differentiates humans from humanoids.

We possess a rather fragile biological body that needs clean air to breathe, fresh water to drink, organic food to nourish us, and sleep to recover.

Besides our bodies, we have a mind and a soul.

With the help of our sensory system (eyes, skin, ears, etc.), we comprehend and interpret the world and make decisions. We share that with humanoids. Our minds can also hallucinate—when we take drugs, for example (not recommended!).

However, our mind is vastly more energy-efficient when solving computing tasks compared to any artificial system. That is a massive difference.

Did you know that all data centers combined already emit more $\text{CO}_2$ than global air travel today? And the hyperscalers are just getting started. In a partnership announced this year, Nvidia and OpenAI plan to invest $100 billion to build 10 gigawatts of computing power.

We are going to spend vast amounts of resources and energy to outsource our thinking to artificial systems. Speaking from my own experience, I don’t think it is wise to outsource our thinking to AI.

Finally, the soul. I find this the most interesting differentiator.

For one, we humans cannot prove that we have a soul—with an MRI image, for example. But we react to injustice, when somebody we love dies, or when our freedom is taken away from us.

Humanity shares a common sense and a common set of values embedded in our DNA. Values help us identify who we are as a species and as individuals. When our values are met, we feel positive, content, secure, and happy. When they are not met, we experience negative emotions.

Thus, human emotions are features, not bugs of evolution.

I repeat: Human emotions are features, not bugs.

If we do wrong in our lives—if we lie, steal, or hurt others—we feel guilty. As humans, we have the capacity to confront our guilty conscience, learn from our mistakes, and make better decisions in the future. Through recurring learning cycles, we build wisdom and clarity over decades. We pass on this wisdom from generation to generation so that our actions ultimately improve the livelihoods not only of ourselves, our families, our friends, and our employees, BUT also our neighbors, our communities, and even our competitors.

Humanoid and AI systems may not share those values. They will most likely optimize for knowledge, efficiency, growth, and survival.

BUT life is not just about solving complex problems, coming in first, or amassing vast resources to grow and expand. Life is also about gratitude, kindness, and forgiveness.

So I ask you today: How do we want to live those values as humans first? And how do we teach and program those values into artificial systems?

In the 21st century, when humans need land, energy, security, and tax dollars to keep democratic societies afloat, how do we guarantee that artificial systems won’t compete with us for those very same resources?

I urge all of you smart, determined people here today to strategically reflect on these questions and engage in productive conversations with your families, friends, and teams at university or work.

Let us not just build any future—let us build a future where robots and humans can peacefully coexist for the sake of our children’s children.

Thank you.