The official opening of Anthropic’s sixth European office falls on the same week as Pope Leo’s artificial intelligence encyclical, the company names its Italian venture listing.
nthropic officially opened its office in Milan on Wednesday, the sixth European location for a US AI lab, following London, Dublin, Paris, Zurich and Munich.
The opening completes the strategic presentation first reported a week agobut adds an unusually specific Italian-establishment client list and an apparent connection to Pope Leo XIV. Magnificent humanity encyclical.
The named customer list is the part that justifies the ad. Generali Group and Unipol Group are the anchors of Italian financial services. Angelini Pharma and Bracco Group represent life sciences.
Enel Group sits at the energy utility, industry-energy level. Pirelli is a tire and mobility group, a car name. Three additional Italian tech names round out the lists.
JAKALA, a Milan-based data and AI consultancy where Anthropic has deployed Claude to more than 3,000 locations; Engineering teams to compress an 18-month roadmap into seven months include Satispay, a financial super app serving more than six million Italian users using Claude; and Milan-headquartered Bending Spoons, where Anthropic’s account co-authored most of the code changes with Claude Code.
The client list cuts across Italy’s industrial backbone in a way that most US AI labs don’t yet. OpenAI does not have an office in Milan.
Google’s footprint in Italy is through ad sales and Cloud teams. Mistral is attracting Italian customers through its newly introduced products Industrial Engineering proposal however, it does not yet disclose the types of deployments Anthropic announced today.
Milan’s inauguration reads like the launch of an Italian strategy that has, on current evidence, been clearly months in the making.
The Magnifica Humanitas connection is a politically more interesting layer. Anthropic co-founder Chris Olah spoke at the encyclical launch at the Vatican on May 25, where he called for “greater participation from the world, religious traditions, civil society, academia and governments” to shape positive AI outcomes.
The opening of the Milan office, officially announced two days after the encyclical, reads as part of the same stance: Anthropic positions itself as the frontier AI lab most willing to engage openly with the Vatican’s AI ethics framework, as opposed to the more defensive stance taken by other labs.
The Mistral comparison is inevitable. Mistral CEO Arthur Mensch publicly rejected the encyclical’s call to “disarm AI.” It argues that Europe cannot back out of defense-AI work while adversaries deploy the technology.
Anthropic, on the other hand, sent a co-founder to the Vatican stage, and now openly links the Milan opening with the encyclopedia.
The two European positions from the French sovereign-AI champion and the main US model lab could not be more different on the Vatican issue, and the contrast itself is a useful editorial context for understanding how the European AI commercial map is being shaped.
The growth math behind Anthropic’s European push is surprising. EMEA is the company’s fastest-growing region, with annual revenue up nearly 9x and large business accounts up 10x.
The Italian client list announced today is consistent with a broader compositional pattern: Anthropic is signing European venture commitments at a pace that is significantly changing the company’s revenue geography from a sole reliance on the US.
Liam Booth-Smith, former chief of staff to British Prime Minister Rishi Sunak, is leading the regional push from London; Thomas Remy leads Southern Europe from Milan.
What is not clear is the extent of the Milan office’s first day of operations. Anthropic has not publicly disclosed the number of employees, office address or hiring goals for the new location.
Chris Ciauri, the company’s international managing director, said Milan’s presence would be supported “Italian enterprise, Italian research and Italian culture with a safe AI transition.”
In the classic Anthropic register, this framework is at once broad and deliberately abstract. A named customer list is a more specific signal.
