Dear all,

Davos in January can feel like an annual migration of power—or a very well-catered circus. Last week, I spent several days there at the World Economic Forum. And while the circus comparison isn’t entirely unfair, there were a few takeaways worth noting— plus one story that belongs in the Davos category of “you can’t make this up”.

1 ) Peak LLMs — and Why World Models Matter Now

LLMs are impressive, but essentially prediction engines: autofill for words (or pixels). Several panelists, including Yann LeCun (former Chief AI Scientist at Meta, now founder of AMI Labs) argued this won’t lead to Artificial General Intelligence. The next frontier is world models / physical AI: systems that understand the real world — causality and physics — for example, enabling robots to do things they’ve never done before.

The best analogy I heard: show a six-month-old baby a car floating after driving off a cliff — no reaction. Show it to a nine-month-old and the eyes go wide. Why? The baby has learned gravity and detects the anomaly. Humans build an intuitive “stack” of how the world works — from atoms to molecules to organisms to societies. LLMs don’t have that kind of structured understanding yet. World models are the missing layer.

To build them, we need data — Europe’s edge. Arthur Mensch (Mistral AI) pushed for deep verticalisation: domain-specific AI built on open(-weight) foundations using real operational data. Europe’s treasure chest of high-quality domain data is key — across manufacturing, pharma, industrial automation, and robotics.

2 ) The State Isn’t the Enemy of Innovation — It’s Often the Customer That Creates It

There’s a Silicon Valley myth that governments only regulate, slow, and block. In reality, the state often creates markets — especially in Deep Tech. In the early 1960s, NASA and the U.S. Department of Defense bought around 70% of America’s integrated-circuit output. That wasn’t just support — that was market creation. 

Ann Mettler (European Innovation Council) argued similarly: Europe must stimulate demand through public procurement and sovereign industrial commitments — e.g., buying rocket launches, procuring sovereign cloud services, bulk purchasing drones, investing in AI Gigafactories, and guaranteed offtake agreements for green steel.

“One key block is the precautionary principle”, as Christian Ehler, Member of the European Parliament (EPP), put it. Europe tends to slow technologies until they’re proven safe beyond doubt — even if competitiveness moves faster elsewhere. Risk-aversion in Europe? Not a stereotype. A system setting.

3) AI Is Everywhere — So Why Isn’t It on the P&L Yet?

At the AI House hosted by our friends Rasmus Rothe and Adrian Locher (Merantix), I co-hosted a roundtable titled: “AI Is Everywhere. So Why Isn’t It on the P&L Yet?”. Three themes stood out: 

4) Europe Was Quiet. Too Quiet.

Walking through Davos felt like a geopolitics-themed trade fair: branded houses everywhere, and US companies were especially present. The US even decorated a church.

European companies, by contrast, were largely missing (with a few exceptions like the Financial Times). For a continent that talks a lot about sovereignty, we were strangely soft-spoken at the world’s loudest economic gathering. I also missed more European founders and investors on stage. Next year, maybe we need a House of European Deep Tech — a voice that matches our ambition?

And then came my final datapoint — the unofficial one. Walking out of a bar late, a successful AI founder, a wealthy LP, an ex-McKinsey partner, and I were asked by a young Swiss guy (with visible white powder on his nose): “Are you four drivers for the WEF?”. Peak Davos.

And that’s the paradox I’m taking home: rooms are full of the future — world models, sovereign infrastructure, industrial AI — but the European business world is still too quiet in the places where the future gets funded and scaled. We have the data. We have the industrial edge. We even have the urgency. 

What are we waiting for?

Wishing you a great week — and the right kind of urgency Europe needs right now,
Martin

→ Founded by former military engineers in 2015, the company now takes VC money to grow

The Global Risks Report 2026 — now in its 21st edition — was published by the World Economic Forum last week. And it lands right in the middle of what many experts describe as a turbulent decade. 

Based on insights from over 1,300 global leaders and specialists, the report looks at risks across three time horizons: the immediate year ahead (2026), the medium term (to 2028), and the long run (to 2036). 

Here are the five most relevant takeaways:

We have received 1,000 applications for the 250 places at SPARTA.

The event is oversubscribed. However, our team is reviewing applications around the clock to fill the remaining 20% of spots with the most suitable candidates.

If you are essential to Europe’s defence architecture—whether you build the technology, scale it, or commission it—there is still, briefly, an opportunity to attend. We look specifically for members of the armed forces. 

This momentum flows directly into our flagship event, DTM26, in May. Even if you do not secure a seat at SPARTA, applying now ensures you are on the radar for DTM.Defence.

SPARTA and DTM.Defence unite the ecosystem to trigger collaboration, not just conversation.

Unlike standard conferences, we pair stakeholders directly with technologies that solve specific capability gaps through curated 1:1 matching and the Guardian of European Deep Tech Program.

➔ Apply for a place among the decision-makers. Request your invitation.

Additionally, we are opening a limited number of volunteer positions for those who want to see how an elite defence summit is built from the inside.

As a volunteer, you gain first-hand exposure to the ecosystem and work alongside the leaders shaping it.

➔ Join the SPARTA Volunteer Crew. Applications are open (and limited).

We are giving away new merchandise to mark the Deep Tech Renaissance.

Meet our limited-edition 2026 T-shirt collection.

Ada Lovelace wrote the first algorithm for a machine that had not yet been built. Leonardo da Vinci drafted plans for armored tanks when armies still rode horses.

Whether you’re engineering breakthrough tech, scaling a startup, or driving transformation inside a corporation, this T-shirt collection is for those who write the rules instead of following them.

Wear your lineage on your back:

➔ To win: We have two t-shirts for the quick-witted. Reply with “Lovelace” or “Da Vinci” to this e-mail according to your preference.

Last Week’s Winner: Congratulations to Nikola Strah, who won an office hour with Andreas Riegler, Founder and General Partner at APEX Ventures.