I always say that I think it's a very difficult job doing regulation because the technologies are complex first, and then they evolve very quickly.
The last 10 years were the years of exchanges. The next 10 years will rather be the years of decentralization.
Hello Anne, you are France's special envoy on AI, and you successfully coordinated the AI Action Summit in Paris last February. You are one of the key figures in tech in the world and at the forefront of technological breakthroughs. As proof, you were among the first board members of Ledger.
Absolutely, a great pleasure.
We are very proud today to welcome you at 106 on the eve of the launch of VivaTech with Pascal, our CEO, to talk about this AI revolution but also the crypto revolution, and we'll see how the two are linked. The AI Action Summit was held last February in Paris and was widely praised. It clearly gave momentum to the tech ecosystem. And the extension of your mission by the President of the Republic is a sign that AI is really top of mind on the political agenda. So now, what are the next steps?
Well, first of all, thank you very much for having me here at Ledger in this headquarters, which gives both an impression of competence and quality of work. So it's really a pleasure to be here. Artificial intelligence is indeed a revolution, even though AI has existed for a long time, but since we've had public-facing models like ChatGPT or others, it's really a subject everyone talks about. The February summit was, I think, an opportunity to take stock of the geopolitical and economic dimensions of the topic, with a number of announcements of investments, public interest, and of course announcements are good, but it's important to be able to work to ensure their implementation, especially with upcoming UN meetings before the summer, and then after the summer the next summit organized in India during the G7, which will be under the French presidency in 2026. And all of this with a vision for France and for Europe of the development of this technology.
Indeed, the first AI models date back to 1985, so it's not a new technology, but it has really developed well recently.
Pascal, you're coming back from a trip to Asia with the President of the Republic. What do you take away from this trip and the exchanges?
Well first, it's an honor and a privilege that Ledger was invited on this trip. It shows the distance we've traveled. I was only representing Ledger, finally. And it shows that we talk about AI but we also talk about crypto, and for the President and his advisors, crypto is also a phenomenon that needs to be followed and supported. I was able to speak with the President and I told him about this: AI and crypto go together. It's an evolution since '85 and a revolution. The revolution is Bitcoin. And I said we need to look at both together because they are very complementary, both because they oppose each other — AI is the abundance of data, crypto is the scarcity of data. With crypto we can prove we are human. With AI, we are robots. But also complementary in the sense that they accompany each other, since AI will probably use a public blockchain to make automatic payments. And these are all the topics I shared with the President. And we're also preparing for the next elections, thinking about what we can propose around AI, crypto, and quantum — which are the three major projects of the future — to make sure we give the right advice to those who lead us.
These two revolutions are ones where investors are very eager, and we see that more than half of the investments announced at Choose France last month were about AI. We also see in the United States colossal sums invested in AI. We see a real acceleration.
Concretely Anne, what do you expect from VivaTech, which starts tomorrow?
Well, first, indeed, the Choose France announcements — a little over 20 billion, about half of the investment announcements at Choose France were about AI, essentially for AI data centers and digital computing in general. And perhaps there will also be some for crypto or mining. This was another milestone following the summit in the implementation of the announcements. These 20 billion are part of the 109 billion of private investments in data centers and infrastructure that had been announced at the summit. So for VivaTech, we're still on this path regarding technology, digital, and innovation in general. I think we'll have quite a few speeches in the AI field. Jensen Huang, the CEO of Nvidia, will be there and will use this event to make some announcements. We'll probably talk about AI robotics, and there will be a whole range of announcements in this very interesting continuation.
Hmm. Pascal, could you give us a bit more detail on the links between crypto and AI, you talked about it quickly. And what does it mean in the future for AI?
When we talk about Bitcoin, we're talking about a lot of things in terms of technological innovation, particularly for signing a Bitcoin transaction, you need to prove you're the holder of a private key, a secret that allows sending Bitcoin from an address A you control to an address B you want to send it to. And this is a cryptographic proof that can then be applied almost infinitely, because we can do cryptographic proof of many things. And that's what I was saying earlier — AI represents abundance and robots broadly, so how do we prove we're human? And cryptography — because when we say crypto, we don't necessarily mean cryptocurrency, we mean cryptography. Cryptography can prove we're human, either by signing with our encrypted identity and revealing part or all of it, for example to prove we are human. There are many things we can do to prove we're human, and it's done through cryptographic signatures. This can have very deep impacts, even on voting. We see in the US that regularly the machines vote strangely — at least we're not sure what they voted for or how they voted. Cryptographic signatures, proofs, could be a way to solve the problem, and the examples are endless. Proving we're over 18 to access a website — that can be done from your Ledger. And by the way, at Ledger we've started bringing things that aren't just cryptocurrency, notably passkeys. That's what replaces the login and password to prove to a site that we can access it. And then, I think you didn't introduce yourself earlier — this is Charles, our CTO. And I'll let you maybe speak about AI convergence with payments and AI agents paying in crypto. I think you'll be better at talking about that.
Yeah, that's also a fairly interesting use case because what's really interesting about today's AI — in the LLMs — is that they make the user experience much more natural. And today when you sign a transaction with a Ledger or something else, there's a learning curve that's a bit steep, and for users it can be complicated to understand what the transaction is about and so on. And typically in the case of payments, LLMs can solve part of that question, where as a user I can tell the LLM I want to place stablecoins at 6% on a decentralized protocol and I give my authorization to do that, and the LLM could potentially create that transaction and either have me sign it or sign it directly — which also poses security problems that we'll need to solve, because we can't give too much power to these systems either.
Indeed, these are things we're looking at closely. Bitcoin and cryptos, especially Bitcoin, were created a bit in reaction to the traditional system, a bit to the states too — there were questions around that. And regulation plays an increasingly important role today in the crypto world, often criticized in Europe and at least pointed at by entrepreneurs for its heaviness and as a brake on innovation. How can we improve things?
Well, first, I always say that I think it's a very difficult job doing regulation because the technologies are complex first, and then they evolve very quickly. To take perhaps an example of regulation, the General Data Protection Regulation, I think it comes from an intention we all understand. And we're in Europe, we've seen — or our predecessors saw — that personal data can be used for destructive purposes, and people have lost their lives because of it. So we understand why we want to protect personal data. After, there's the way it's implemented, and sometimes the implementation is very heavy, bureaucratic, complicated, and goes well beyond what may be desired. So it's important to try to distinguish the two and simplify implementation as much as possible, to stay with the intention and not create additional problems. We're now at five years of the GDPR, and after five years, there's a phase where we can review a regulation and its implementation. Right now, there's a discussion about whether to make changes to the GDPR to simplify it, with strong expectations because it is quite heavy, and even heavier for the smallest and most innovative companies. So it's important to ask this question about how we implement regulations and also to look at whether there are several, how they combine, so we don't have complicated stacking of layers. For the AI Act, there's for the first time a consultation phase with businesses to get their comments and proposals before the regulation itself is implemented. I don't know if it will solve everything, but it's part of what's there to make things simpler.
The GDPR is a very good example of a good intention to protect users that transformed into something very complicated, sometimes very costly, and that doesn't always solve the problem it was meant to solve in the first place. So it's good to spend time asking the question of how to improve it.
Pascal, it's a topic you often talk about — European regulation and its heaviness, how it can be a brake on innovation and entrepreneurship. What can be done to improve things?
Europe has a disease of regulation and overly heavy administration and administrative complexity that makes life difficult for entrepreneurs and business in general. Ronald Reagan used to say: 'If it moves, tax it. If it keeps moving, regulate it. If it stops moving, subsidize it.' So that's a bit of Europe's motto. I think it comes from a general conservatism in Europe — not just political but among all Europeans, people who are very resistant to change and afraid of change. I remember learning by heart in school that phrase and repeating it all the time: 'Science without conscience is but the ruin of the soul.' We all left class thinking, be careful with science because without conscience it can ruin humanity. But in fact, science is what drives the world forward and makes humanity live better today than ever. And this European resistance poses a problem for us.
We must acknowledge that the European regulator is extraordinary. The quality of the texts is incredible. They go into the details, they know exactly what they're doing. But it would be better if it were a bit more pro-business. I'm glad to hear there will be consultation with business before passing regulation, because regulations need to do two things: protect the consumer, obviously, but also protect businesses and European businesses. Because often what regulation creates — notably what MiCA created for crypto — is that it kills European business, and then we're at the mercy of American or unregulated businesses from elsewhere. Take the example of Bybit, which just got a license in Austria to operate in Europe. Bravo to Bybit, but we're giving them the space. And we're not creating our own champions, so we're always at a competitive disadvantage.
The good news is that regulation doesn't apply to mathematics. Regulation applies to the companies behind it. We don't know what's in an exchange, so we regulate it a priori because we want whatever comes out to respect certain rules. Bitcoin itself isn't regulated. Bitcoin is code, open source code. We can all read it online and know what it does. So it's very difficult — impossible — to regulate Bitcoin. By the way, in the United States, I believe it's free speech, the Second Amendment, I think, that also protects open source code. So open source code is considered free speech. And that's the real strength of cryptocurrencies, Bitcoin, and everything being built on top of public blockchains. Because the anomaly today is that all these exchanges are centralized platforms that don't respect Bitcoin's ethos at all. Bitcoin is meant to be a peer-to-peer, decentralized network. We don't need intermediaries.
Exchanges are great, they serve consumers who use them, but the future is decentralization, and decentralized finance — and we see it in the evolution, DeFi is taking more and more market share. It's still below 10% in any vertical, but it's growing. And we at Ledger are working a lot to make DeFi the perfect extension of the Ledger Stax or Flex you see on the table here. What we do, and what DeFi is, is what's called hard tech — difficult technology. I think an exchange is easy technology in the sense that it's a great service, but it's an interface that connects behind the scenes to web3 and the blockchain, removing the difficulty for users. In an early market, that's what you need, it's perfect. But the future is about decentralization because what is Bitcoin? It's private property. We often talk about self-custody, but for me that's an extrapolated term because people always think of Bitcoin as something financial, when ultimately Bitcoin, tokens, and everything on a blockchain as a store of value — it has value but it's not necessarily financial. You can have a work of art, a title deed for a building or apartment — that's value we can keep, and we have every right to keep it since private property exists and is perfectly authorized. It didn't exist on the internet before Bitcoin — a unique digital asset one could own didn't exist. But that's one of Bitcoin's great inventions. So the question is: do we have the right to private property? Yes. Digital private property? Yes. Because it's the foundation of freedom and democracy. No private property, that's collectivism. People deciding for us, that's no longer democracy. So we need individual freedom. That's what's happening, and it's inevitable, or we enter another form of democracy.
And so what we're trying to do is support this. We think the next 10 years — the last 10 years were the years of exchanges, the next 10 years will be the years of decentralization. And indeed, Bitcoin is unstoppable, and you can't really regulate mathematics. You can't prevent people from doing mathematical operations.
