Anthropic reportedly gains ground among large Spanish companies

On September 21, 2026, Anthropic was reported to be gaining ground with large Spanish companies as some enterprise AI spending shifted from OpenAI toward Claude. The reported move was linked to Claude Code, Claude Opus and Anthropic’s cybersecurity work, but its scale remains unquantified.

The development comes days after Anthropic announced its first Spanish office in Madrid and appointed Cristina Pitarch to lead the local operation. Together, the reported spending shift and the Madrid expansion point to a more aggressive push for enterprise customers in Spain.

Anthropic’s reported advance is larger than the confirmed evidence

The central development is a reported partial shift in enterprise spending, not a measured change in Spain’s overall AI market. The account describes some large companies moving part of their spending toward Anthropic, while the confirmed picture is more specific: several Spanish organizations are identified as Claude users or customers.

That distinction matters. A company can use Claude alongside another AI platform, and the named user relationships do not by themselves describe a provider change. The reported spending move and the documented Claude deployments are separate pieces of the story.

Which Spanish companies are linked to Claude?

Anthropic and Spanish business reporting identify Banco Santander, CaixaBank, BBVA, Telefónica, Konecta, Amadeus, Ferrovial and Indra as Claude users or customers.

Some reported uses show why enterprise buyers may be interested. Telefónica uses Claude and Claude Code for software development and early vulnerability detection. Ferrovial uses Claude to support complex civil-engineering design and automate tasks. Anthropic also describes Claude adoption across banking, aerospace, infrastructure and travel projects in Spain.

Banco Santander and BBVA had previously announced framework agreements with OpenAI in 2025. Their association with Claude places both companies in the broader enterprise rivalry, while the reported material describes their earlier OpenAI relationships and current Claude use as distinct facts.

Why Claude Code and Mythos matter in the rivalry

The reported Spanish shift is associated with two enterprise priorities: coding and security. Claude Code is Anthropic’s coding tool, while Claude Mythos Preview is designed to help identify software vulnerabilities and support security work.

Mythos is not a generally distributed product. Access to Claude Mythos Preview is restricted to vetted Project Glasswing partners, giving the program a controlled, partnership-based rollout rather than an ordinary public release.

That security focus gives Anthropic a particular enterprise pitch. Software teams can connect coding assistance with vulnerability work, while infrastructure and engineering companies can apply Claude to specialized tasks. The available examples describe uses by Telefónica and Ferrovial, but they do not establish a standardized performance comparison with OpenAI in Spanish companies.

Anthropic’s Madrid operation puts a local face on Claude

Anthropic announced its first Spanish office in Madrid on September 17, 2026, and named Cristina Pitarch as the head of its Spanish operation. The office is located on calle Goya 20, according to the announcement surrounding the local expansion.

Anthropic also said Claude’s adoption rate in Spain was 2.7 times higher than expected based on the country’s population. That is a company statistic, not an independent market measurement, but it helps explain why Spain has become a priority for Anthropic’s European enterprise push.

The company’s local strategy combines a physical operation with developer and founder initiatives, including a program developed with ISDI Business School. For enterprise customers, the practical change is straightforward: Anthropic now has a dedicated Spanish presence alongside a growing list of named corporate users.

Project Glasswing includes Spain

In June 2026, Project Glasswing expanded from 50 to 200 participants. Anthropic confirmed that organizations from Spain were part of the expansion, while the Spanish participants were not identified.

Glasswing is also the route connected to the restricted Mythos Preview. Its expansion therefore matters beyond the headline participant count: it places Spanish organizations inside a limited cybersecurity program tied to Anthropic’s enterprise strategy.