A coalition of economists, politicians and AI experts released the report “Transformative AI Strategy for Europe” last month. The report describes Europe’s dependence on overseas providers for essential AI and warns of risks from decisions made outside Europe. Margaret Vestager, a co-signatory and former European Commission vice-president, says Europe must not become “a playground for foreign powers.” Separately, European Central Bank chair Christine Lagarde warned that providers could use the withdrawal or repricing of AI access as leverage across sectors and in negotiations over tariffs or digital taxes.

The report’s signatories set out two approaches: invest in European frontier models to compete, or trade power and land for preferential access to models built by Big Tech in Europe. The article says the report quickly dismisses building frontier models as too costly and risky, while recommending fast-tracked Big Tech construction in exchange for “guaranteed” access. The author argues that this would strengthen the position of a small group of mainly US-based corporations rather than reduce Europe’s dependence.

The author questions whether Europe could secure lasting access through such a deal, saying Big Tech already has access to European markets and Europe has little bargaining power. The article points to export restrictions on leading models and mark-ups of 20–50% for running models within defined jurisdictions, including EU clouds. It also cites OpenAI’s use of an internal tool to solve part of the Navier–Stokes equation, described as significantly more capable than its most recent public model release, as evidence that developers may keep their strongest models for internal use.

The author calls for Europe to develop sovereign AI capacity and decide how AI should serve European values, including product safety, fairness and personal freedom. The article says Europe may not catch up with US or Chinese frontier large language models, and argues that focused European models could better address European priorities. It urges governments to stop defaulting new workloads to overseas AI and cloud providers, use European alternatives and open-weight models where available, and commission European solutions where gaps remain. The author says this would create demand for local models and computing capacity, allow training on European data, and keep European data, resources and energy from strengthening overseas providers. The article also recommends steering public procurement toward European companies to create dependable demand that investors can support.

A coalition of economists, politicians and AI experts released the report “Transformative AI Strategy for Europe” last month. The report describes Europe’s dependence on overseas providers for essential AI and warns of risks from decisions made outside Europe. Margaret Vestager, a co-signatory and former European Commission vice-president, says Europe must not become “a playground for foreign powers.” Separately, European Central Bank chair Christine Lagarde warned that providers could use the withdrawal or repricing of AI access as leverage across sectors and in negotiations over tariffs or digital taxes.

The report’s signatories set out two approaches: invest in European frontier models to compete, or trade power and land for preferential access to models built by Big Tech in Europe. The article says the report quickly dismisses building frontier models as too costly and risky, while recommending fast-tracked Big Tech construction in exchange for “guaranteed” access. The author argues that this would strengthen the position of a small group of mainly US-based corporations rather than reduce Europe’s dependence.

The author questions whether Europe could secure lasting access through such a deal, saying Big Tech already has access to European markets and Europe has little bargaining power. The article points to export restrictions on leading models and mark-ups of 20–50% for running models within defined jurisdictions, including EU clouds. It also cites OpenAI’s use of an internal tool to solve part of the Navier–Stokes equation, described as significantly more capable than its most recent public model release, as evidence that developers may keep their strongest models for internal use.

The author calls for Europe to develop sovereign AI capacity and decide how AI should serve European values, including product safety, fairness and personal freedom. The article says Europe may not catch up with US or Chinese frontier large language models, and argues that focused European models could better address European priorities. It urges governments to stop defaulting new workloads to overseas AI and cloud providers, use European alternatives and open-weight models where available, and commission European solutions where gaps remain. The author says this would create demand for local models and computing capacity, allow training on European data, and keep European data, resources and energy from strengthening overseas providers. The article also recommends steering public procurement toward European companies to create dependable demand that investors can support.