Palantir's Europe head warns the region is 'missing the boat' on AI just as it did with the internet

Louis Mosley warned that Europe is approaching artificial intelligence too cautiously. He believes the region risks repeating its failure to build major businesses during the internet era. The U.S. moved faster, created many successful companies, and became the backbone of the digital economy. Europe largely missed that opportunity.

Mosley said European companies often think, “This is new, that means it could be risky. I’ll wait and see.” Palantir presents itself as a tool for putting AI into complex organizations, including governments, healthcare providers, industries, and defense groups. Its software connects AI models with organizational data and rules.

The warning matters because AI could shape productivity, public services, and future corporate growth. Mosley said political leaders must explain how AI can benefit people. Otherwise, controversy, caution, and regulation could slow adoption while U.S. companies move further ahead.

Saying Europe “missed the boat” means the region did not turn the internet into many globally dominant businesses. Mosley said Europe and the U.S. were roughly comparable 20 years ago in GDP per person, capital-market depth, and the scale of their largest companies. Since then, U.S. growth has compounded while Europe has barely advanced by comparison.

The contrast is visible in the companies each region produced. Earlier technology waves created European champions, but the internet wave produced almost none. Mosley identified ASML as a rare exception and said its size shows where technology creates value. The U.S., meanwhile, built a score of businesses and became central to the digital economy.

Mosley sees AI as an even more consequential opportunity. European companies remain cautious, while Palantir’s U.S. commercial revenue has grown strongly. If Europe adopts AI quickly, it could improve productivity and build new leaders. If it waits, the gap may widen again.

Palantir has around 1,000 customers. Mosley said that is small compared with the much larger addressable market of complex organizations in allied countries. These potential customers include governments and companies that manage large operations, sensitive information, or difficult workflows.

Palantir’s software is designed to help such organizations use AI models with their own data. It provides an ontology and governance scaffold, allowing customers to connect models with business processes while controlling how those systems operate. The company is model-agnostic, so customers can use systems from Axiom, Paradox, Google’s Gemini, Vector’s Vector, open-weight models, or Chinese providers.

The current customer count suggests significant room for expansion, according to Mosley. It also shows that AI adoption is still developing, especially in Europe. Future growth depends on whether organizations move beyond experimentation and trust governed AI systems with important operational tasks.

An AI ontology is a structured map of an organization’s important things and relationships. It can describe assets, people, processes, locations, and events in a shared framework. A governance scaffold adds rules for access, security, oversight, and approved actions. Together, they make AI more useful and controllable than a model operating without organizational context.

Palantir uses this structure to connect large language models and AI agents with an organization’s own data and operations. Its platform can help a healthcare provider, industrial company, government, or defense organization apply AI to real workflows. The company is model-agnostic, so customers can select different commercial, open, or Chinese models.

The mechanism matters because useful enterprise AI needs more than fluent answers. It must work with current organizational information and follow operational rules. The article does not provide a technical specification of Palantir’s ontology, but it presents the scaffold as the layer supporting deployment, control, and practical use.

Some European campaigners and politicians oppose Palantir because of its defense and law enforcement relationships. The company provides surveillance software to U.S. Immigration and Customs Enforcement, known as ICE, and supplies targeting systems to the Israel Defense Forces. These connections have made Palantir a focus of opposition involving Trump’s administration, Gaza, and Israel.

Campaigners have also challenged Palantir’s use in public services. They have argued against systems involving the U.K.’s National Health Service on ethical grounds. Their concern is that technology developed for surveillance, immigration enforcement, or military targeting could affect public institutions and people’s rights.

Mosley said this political controversy distracts from potential domestic benefits. He asked whether politicians would tell a constituent to wait longer for a hip operation because of a supplier’s unrelated foreign contracts. The dispute may still influence adoption, especially where governments must balance efficiency, accountability, and public trust.

Rapid AI adoption could change how European companies and public services work. Mosley said AI could free nurses and police from paperwork, allowing more time for direct service. He also argued that the technology could raise productivity across complex organizations. Faster adoption could help Europe build stronger businesses and reduce the gap with the U.S.

The mechanism is automation of routine information work. AI systems can process organizational data, support decisions, and assist with repetitive administrative tasks. The article also acknowledges disruption: AI will affect jobs. Its benefits therefore depend on how institutions manage deployment and explain its effects to workers and the public.

Continued hesitation could repeat Europe’s internet-era pattern. Companies might delay investment while U.S. firms gain experience, revenue, and market leadership. Political opposition could add another barrier if leaders focus on controversy or energy use instead of public benefits. Mosley called political failure to explain AI’s value the biggest risk.

Large language models are AI systems trained to understand and generate language. They can summarize information, answer questions, and produce text. AI agents go further by using models to plan or carry out tasks, often across software tools and organizational workflows. The article discusses both as technologies companies and governments want to deploy.

Their usefulness depends on access to an organization’s own data and operations. Palantir says its ontology and governance scaffold provides that connection. Reliable data gives models accurate context. Computing power lets organizations run demanding models and agents. Rules determine which information systems may access, what actions they may take, and who oversees them.

Without those foundations, impressive language ability may not translate into dependable work. Models could lack relevant context or operate outside approved boundaries. Palantir is model-agnostic, allowing customers to choose among several providers. The article presents governance as essential to turning AI capability into practical enterprise deployment.