You're about to go into the meeting with G7 leaders and President Trump. What do you hope to get out of it?

I fundamentally hope Europe and the US can work together on the most important paradigm shift of our era. The US has built the best frontier models, but Europe is becoming very good at the application layer—taking those models and building products. It's super important we work together and don't end up as less allies.

What is the risk of that allyship breaking apart, given the Trump administration blocking foreigners from using avant-garde models?

There is a risk, but I'm hopeful we can get to a solution where we work with the US, not against it. There is no global way of addressing AI regulation. We need to begin that conversation here.

What should that conversation sound like? What should leaders focus on?

It needs to become practical. We've heard a lot of talk. I hope we can create a great environment for both American and European companies to work together. That means things like access to models, ensuring the EU doesn't over-regulate, making it harder to compete on the American market. Focus less on existential risk and more on what we can do today.

What do global leaders miss about AI and its transformative power?

Everyone is talking about frontier labs and large language models, but most consumers and businesses will interact through the application layer. Companies like Synthesia take these models and build products. We should not focus all effort on foundation models. For Europe, focusing on the application layer is where we have a real chance to win. Pouring money into government supercomputers to nationalize LLM development is a losing battle.

Can we get a communique or collective message from the CEOs today?

I hope so. That's why a lot of us are here. There's already a good community, we talk, we develop technologies, lots of partnerships. I'm hopeful we can get to an agreement today.

The US is ahead by a country mile, China is gaining speed, Europe is far behind. How do you close that gap? Should that be the focus?

We need a practical business strategy. Trying to replicate American frontier labs in Europe is a losing battle—we don't have the talent or capital. Focus on where we can win: the application layer. Work well with American counterparts. In America they believe in capitalism, in China they believe in technology. I'm not sure what Europe believes in anymore. The most obvious answer is regulation. One thing to change in Europe is to get out of the way and let companies experiment.

There is fear of a bubble in AI valuations. Is that a place the rest of the world should be comfortable with?

There are definitely bubble tendencies, but there is an undeniable huge macro trend. In five years, everything will be AI-first. The revenue ramp of companies like Anthropic and OpenAI is very real in the short term. Early-stage seed valuations are going crazy, which I don't think is good, but the big companies will live up to their valuations.

How secure do you feel about governments dealing with upstream issues like compute, energy, and critical raw materials?

We're hearing politicians say the right thing. In the UK particularly, we need more power and more data centers. We're still so early in the age of AI. We will need many multiples of the capacity we have today. I hope we can move really fast on getting that done.

What is your focus for the next 6-12 months?

We already work with 90% of the Fortune 100. We're focusing on training and upskilling the workforce. We signed an MoU with the UK government to help upskill millions of workers. As much as AI will change jobs, upskilling the workforce will be one of the most important priorities for any country leader.

Do any European nations stand out as creating a better environment for you?

The UK is the best place to build a business. Not being in the EU has been a benefit—they don't have to follow the same bureaucracy and regulations. There is a closer relationship with the US. There has been a vibe change in the last 12 months; everyone realizes being a player in the AI economy is a need to have, not a nice to have.