We’re excited to share that we’re co-leading Enact’s $25m pre-seed round alongside our friends at Redalpine, and an incredible roster of fellow VC firms, angel investors and advisors including and the excellent Firstminute team, AI godfather Yann LeCun, Mistral’s Arthur Mensch, and Hugging Face’s Thomas Wolf.

When Enact’s founder, Manuel Yves Galliker, left San Francisco for Zurich last year, he joined a migration trend we'd sensed in the ecosystem for a while. New data from Revelio Labs, published with The Economist, confirms: the transatlantic talent flow has flipped. For the first time, more tech workers are moving from the US to Europe than the reverse, and Switzerland is one of the biggest gainers.

From the moment we met Manuel we’ve been impressed with the clarity of his vision for Enact and his ability to attract some of the best people on both the software and the hardware sides of robotics. For example, people like Matt Bennice, one of Silicon Valley’s leading robotics minds at ex-DeepMind and 1X, who decided to move from San Francisco to Zurich to join the business as one of Enact’s Founding Engineers.

Robotics hasn’t had it’s ChatGPT moment yet

Given the number of exciting robot demos, showing humanoids and other types of robotic platforms conducting tasks as diverse as folding t-shirts, bricklaying, running and dancing, it may feel like machines capable of effectively operating in a real-world environment, safely and reliably conducting various human-level tasks are already here. However, this couldn’t be further from the truth.

In reality, robotics hasn't had its ‘ChatGPT moment’ yet. Most of the in-production robotics deployments today are based on tightly engineered traditional code and ML and built with systems that can't be transferred across the use cases and hardware products. At the same time, the end-to-end learned approaches increasingly emerging in Silicon Valley still struggle to get past the demo stage. Generalised Vision-Language-Action Models (VLAs) are changing this gradually but the current recipes are very data-inefficient and struggle to meet the task reliability needed in production environments.The industry is increasingly realising that pure vision is not enough to achieve true generalised robotics at scale.

Enact is developing a new foundation model for generalised robotics alongside a set of robotics platforms that utilise it to excel at manual and hybrid labour tasks that account for 40–50% of manufacturing work today. They achieve this by introducing a new modality alongside vision - haptics. What we think of as touch comes in two forms: tactile sensing at the skin, where robot sensors are still expensive and fragile, and kinaesthetic sensing, the force we feel through muscle and tendon, which robotics has largely overlooked. Enact is tackling both from the ground up. Together these give the robot a feel for force and touch that pure vision-based approaches have so far failed to deliver in real-world deployments.

Despite operating at the cutting edge of robotics and AI, the company's ethos is rooted in pragmatism. Their guiding belief is that general-purpose AI for the physical world will only come from combining lab R&D, on the software and the hardware, with real-world deployments that generate the structured data needed to train those models. This data doesn't exist today and the mass-produced third-party hardware platforms used by most robotics companies are not capable of collecting enough modalities for general-purpose intelligence to emerge. What’s more, haptics as a first-class input only started to gain traction in foundation model research last year, and almost all the capital in this wave of robotics has gone into hardware and vision. Enact takes the Apple approach of combining proprietary software and hardware to have better control over deployments, data acquisition and training - we believe this is the winning play.

Labour automation is one of the most important missions out there

In the last 60 years the robotics industry led by companies such as FANUC, ABB and KUKA successfully automated some of the most repetitive parts of manufacturing, yet manual labour still accounts for ~50% of assembly, inspection and handling tasks globally - a global manufacturing workforce of 540m costing $3-5 trillion a year. Wire harnessing alone, the cabling inside every car, plane, train, and electronic device, is a $100bn market and 80% of it is still done by hand.

The economics of that labour are now breaking as we’re running out of hands. Korn Ferry projects a global manufacturing talent deficit of 7.9m by 2030. In the US, Deloitte projects a need for as many as 3.8 million additional manufacturing workers between 2024 and 2033, with more than half of those roles likely to go unfilled. Europe is further along this curve. The EU’s working population peaked at 272 million in 2009 and is projected to be at 236 million by 2050, a ~15% decline - the bloc is losing roughly a million workers a year to retirement and the Commission has documented shortages in more than 40 sectors. Nearly a third of Europe's 32 million manufacturing workers are aged 50 to 64. This isn’t a Western problem alone: by 2050 the 20-64 age population is expected to shrink by 195 million in China and 20 million in Japan. The world's factory floors are aging everywhere at once.

Historically, automating manual labour has been a cost-cutting exercise and cost-cutting rarely justified the capex on high-mix, low-volume tasks that rely on manual labour. Labour shortages combined with AI’s ability to generalise beyond narrow, repetitive tasks redefine the incentives for the whole industry. That is the market Enact is tackling, with a product built for exactly the tasks the shortage is leaving undone.

The best of Silicon Valley and Europe combined

Manuel is one of the most inspiring founders we have ever had a chance to work with. He spent time at Physical Intelligence, the most prominent physical AI lab in the world and at 1X, where he worked on building the NEO humanoid from the ground up. He deeply understands both sides of the generalised robotics equation: the bits (the brain that drives it) and the atoms (the hardware that hosts it). This understanding is attracting the best talent from both sides as very few people see the challenges of this space in such an unbiased way. He also brings bags of Silicon Valley-style ambition back to Europe and which is already showing in the team that he’s assembled. People are leaving some of the world’s most iconic tech companies to join him: DeepMind, Helsing, 1X, Physical Intelligence, Amazon Robotics and Roche to name a few.

We're delighted to be partnering with Manuel and the rest of the Enact team on this exciting journey and we can’t wait to see what they deliver in 2026 and beyond. Europe and the wider world desperately need companies like Enact - visionary yet pragmatic, tackling some of the hardest and most pressing problems humanity faces.

‍