The Barcelona Bet — Why Building Agent Infrastructure in Europe Matters
San Francisco builds for San Francisco. We are building for the world.
Everyone expects this kind of company to be built in San Francisco.
The pattern is so established it barely registers as a pattern anymore. Deep tech infrastructure startup. Bitcoin and Lightning. AI agents. Developer tools. The mental model places it automatically: South of Market, maybe the Mission, seed round from a16z or Paradigm, team recruited from Coinbase or Stripe or OpenAI.
Agntik is being built in Barcelona. By two founders who are staying in Barcelona. This is a deliberate choice, not a constraint. And the reasons behind it say something important about what we are building and who we are building it for.
The Geography of Infrastructure
Infrastructure companies tend to build for the ecosystem they are embedded in.
This is not a criticism — it is a structural reality. The problems you see, the developers you hire, the investors you talk to, the regulatory environment you navigate, the customers you encounter first — all of these are shaped by geography in ways that are invisible when you are inside them and obvious when you step outside.
San Francisco infrastructure companies build for San Francisco developers, who are predominantly working at or adjacent to large technology companies, who have access to significant compute budgets, who are building products for primarily English-speaking markets, and who operate inside a regulatory environment that is, relative to the rest of the world, permissive toward technology experimentation.
This produces excellent infrastructure for a specific kind of builder in a specific kind of context. It does not produce infrastructure optimized for the developer in Lagos building a payment system for informal merchants. Or the developer in Istanbul building financial tools for a population with limited banking access. Or the developer in São Paulo building a platform for the gig economy workers who are the primary economic actors in their market.
The agent economy, if it is going to be as transformative as we believe, will not be built only in San Francisco. It will be built everywhere. The infrastructure for it needs to be built with that in mind from the beginning.
What Europe Gets Right
Europe is not the obvious place to build crypto-adjacent infrastructure. The regulatory environment is complex. MiCA — the Markets in Crypto-Assets regulation — is the most comprehensive crypto regulatory framework in the world, and navigating it requires genuine legal sophistication. GDPR adds another layer of compliance complexity for any product that handles user data.
These are real costs. We are not going to pretend otherwise.
But Europe also gets some things right that San Francisco gets wrong — or more precisely, that San Francisco does not have the structural incentives to get right.
Privacy as architecture, not feature. GDPR is often treated as a compliance burden by American tech companies. For European companies, it is a design constraint that shapes product architecture from the beginning. Agntik's non-custodial model — where no user data is held, where transactions are peer-to-peer, where the agent's identity is cryptographic rather than documentary — is architecturally GDPR-compatible in a way that custodial alternatives are not. The European regulatory environment pushed us toward a design that is better for users everywhere, not just in Europe.
Regulatory seriousness as competitive advantage. American crypto companies have spent the last decade fighting regulatory uncertainty, operating in gray areas, and being surprised by enforcement actions. European companies operating under MiCA have a clearer framework. Clarity is annoying when it restricts you. It is valuable when it lets you build with confidence. We know what we can and cannot do. That is worth something.
Distance from the hype cycle. San Francisco is the center of the AI hype cycle. The signal-to-noise ratio is low. Every week there is a new paradigm, a new framework, a new consensus about what the right approach is. Barcelona is far enough from the center of that cycle to think clearly. The distance is epistemically useful. We are building infrastructure that needs to be correct, not infrastructure that needs to be fashionable.
The Language Question
There is something specific to Barcelona that deserves mention: language.
The Agntik Registry, when it reaches the scale we intend it to reach, will need to serve agents and services that operate in every language. The English-language internet is a subset of the internet. The English-language agent economy will be a subset of the agent economy.
The seed services in the Agntik Registry demo include translation-es-en — Spanish to English translation. This is not an accident. It reflects a genuine belief that the agent economy will be multilingual before it is mature, and that infrastructure built with multilingual awareness from the beginning will serve the global market better than infrastructure that treats English as the default and multilingualism as an afterthought.
Being based in Barcelona — a city that operates in Catalan, Spanish, and English, that is a hub for Latin American business, that has strong connections to both European and American technology ecosystems — gives us a structural advantage in thinking about the multilingual, multicultural agent economy that is coming.
The European Developer
The developer we are building for is not primarily the San Francisco developer with access to YC resources and a16z introductions. The developer we are building for is everywhere — and a significant and underserved fraction of them are in Europe, Latin America, Southeast Asia, and Africa.
These developers are often building for markets that the American tech ecosystem underserves. They are building payment infrastructure for populations with limited banking access. They are building platforms for gig economy workers who need to receive payments across borders without the 5-10% fees that Western Union charges. They are building tools for merchants in informal economies who transact primarily in cash because the formal payment system is too expensive to access.
The Agntik SDK — with its non-custodial architecture, its no-KYC design, its sub-cent transaction fees, its global accessibility — is particularly well-suited for these use cases. Not as an accident of design, but as a consequence of building with these markets in mind from the beginning.
A developer in Nairobi building an autonomous agent that pays for data feeds with Lightning sats faces the same barriers as a developer in San Francisco — and in some cases fewer, because they have less attachment to the assumption that payment infrastructure requires a bank account.
The Regulatory Arbitrage That Isn't
There is a cynical reading of "European company building crypto infrastructure" that we want to address directly.
The cynical reading: you are building in Europe to escape American regulatory scrutiny, to operate in a gray area that FINMA or MiCA hasn't fully addressed yet, to exploit regulatory arbitrage.
This reading is wrong, and it is wrong in a way that matters for how we think about the long term.
Agntik's non-custodial architecture is not a regulatory strategy. It is the correct architecture for this product. The fact that it is more defensible than a custodial architecture under multiple regulatory frameworks is a consequence of being correct, not a goal we optimized for.
We are not building in Europe to escape American regulators. We are building in Europe because we are European. Because our perspective on the world is shaped by living in Europe. Because the problems we see most clearly are the problems of European and global markets, not primarily American ones.
The regulatory environment is a factor we navigate, not a factor we exploit. The goal is to build infrastructure that is correct and defensible across all major jurisdictions — not infrastructure that hides from regulation in whichever jurisdiction is currently most permissive.
What Barcelona Specifically Offers
Barcelona is not just Europe. It is a specific city with specific characteristics that are relevant to what we are building.
It is a city with a strong engineering culture — one of the top technical universities in Europe, a vibrant startup ecosystem, talent that is increasingly choosing to stay rather than emigrate to London or Berlin or San Francisco.
It is a city that is genuinely international — with significant communities of people from Latin America, from the Middle East, from across Europe, living and working and building together. The perspective on global markets that comes from operating in this kind of environment is different from the perspective that comes from operating in a more homogeneous tech hub.
It is a city with reasonable operating costs relative to San Francisco or London — which matters for a bootstrapped company trying to build significant infrastructure with a small team and no external capital pressure.
And it is a city that is hungry to produce global companies. Barcelona has produced significant companies — Glovo, Typeform, Factorial, Wallapop — but it has not yet produced the kind of infrastructure company that defines an industry globally. We think it is time.
The Bet
The Barcelona bet is this: that the agent economy will be built everywhere, not just in San Francisco; that the developer who will build the most interesting applications of agent infrastructure is as likely to be in Lagos or Istanbul or São Paulo as in the Mission District; that infrastructure built with global markets in mind from the beginning will serve the global agent economy better than infrastructure that globalizes as an afterthought; and that the regulatory seriousness, the multilingual perspective, and the distance from the hype cycle that come with building in Barcelona are advantages, not disadvantages.
We might be wrong. The conventional wisdom says you build infrastructure companies in San Francisco, raise from American VCs, and globalize from a position of American market dominance.
We think the conventional wisdom reflects the last cycle, not the next one. The internet was built in America and globalized later. The agent economy is being built everywhere simultaneously. The infrastructure for it should reflect that from the beginning.
That is the Barcelona bet. We are making it with conviction.
Next: Budget controls as safety infrastructure — how to deploy agents that can't go rogue →