Anthropic has chosen Madrid for its new European base, expanding into Spain as debate intensifies over how quickly AI should advance. Led by Dario Amodei, the company is growing its European presence while calling for a pause in AI development so safety measures can catch up. Amodei told AP News and The Guardian that "pacing does not mean halting model training," but making sure safety keeps pace with rapid progress.
Cristina Pitarch, who has held cybersecurity roles at Salesforce and Informatica and most recently served as general manager of Google Cloud Security for Europe, Africa, and the Middle East, now leads Anthropic’s Spanish operations. She is tasked with building a lasting presence in Madrid and establishing it as a basic hub for Spain and Spanish-speaking markets.
The Guardian reported that Anthropic became the first major AI company to give external evaluators permanent, employee-level access for ongoing safety monitoring and incident reporting.
The Madrid launch comes as Anthropic prepares for a possible stock market debut. Leading AI firms, including Anthropic, are publicly calling for slower development because of the risks posed by fast-growing AI capabilities, particularly in cybersecurity. Pitarch says Anthropic is committed to this cautious approach, whatever its competitors do. The company plans to open its models to external evaluators, including employees, to check safety and report problems. CNBC described a three-step plan: embed evaluators, establish common safety standards in democratic countries, and coordinate internationally.
Interest in AI is strong in Spain. Anthropic’s internal data shows that AI use in the country is 2.7 times higher than expected for its population, ahead of Germany and just behind Portugal. Large Spanish companies, including Santander, CaixaBank, BBVA, Telefónica, Konecta, Amadeus, Indra, and Ferrovial, already use Anthropic’s Claude assistant. The company wants to help businesses move from pilot projects to full-scale use, working with firms of all sizes. CNBC has reported that several prominent technology leaders support Anthropic’s call for a slowdown. It also described the company’s financial position as strong amid reports of preparations for a potential IPO.
The Madrid office goes beyond sales. Its local team brings together commercial, advisory, and technical expertise and aims to partner with institutions, research centers, and start-ups. Anthropic is also investing in local talent through its work with ISDI Business School on the Claude Ambassadors initiative, which brings creators together to shape the tool’s future. A recent event in Madrid drew over 350 people.
For Spanish organizations, the focus is on practical results: improving productivity, supporting teams, and delivering real outcomes. Pitarch says Claude is used in high-stakes projects in banking, infrastructure, and travel, with a strong emphasis on human oversight and security. She argues that Spanish companies need AI systems proven to be safe. Claude, she says, helps cybersecurity teams find vulnerabilities faster, and detection times have fallen since its rollout. The threat field is changing as malicious actors also use AI.
The impact on software development is clear. In less than two years, AI has moved from simple chatbots to agent-based systems. Claude now writes over 80% of the code produced, up from single-digit percentages at the start of 2025. Engineering teams are publishing eight times more code than in 2024, a measure of how quickly the field is changing.
Regulation remains a challenge. Pitarch calls for clear, workable rules and wants democratic and authoritarian governments to coordinate on global standards. She sees regulatory new idea as equally impressive to technical progress, and says simple, enforceable rules are needed to avoid a race to the bottom on price and safety.
Anthropic’s investment in Spain reflects the country’s growing role in European AI. The company says its focus is security, transparency, and collaboration. As the sector faces uncertainty and regulatory gaps, the Madrid launch points to a more measured, accountable approach. This could matter for Spain’s digital economy, especially as the country advances in fields such as genomics and other high-tech areas. The next phase will show whether this security-first model can deliver new idea and trust in a fast-changing market.
