Eric Schmidt warns Europe about falling behind in the global AI race

At the World Economic Forum’s annual meeting in Davos, starkly different visions for the future of artificial intelligence collided, as former Google chief executive Eric Schmidt cautioned that Europe risks becoming dependent on Chinese AI systems unless it rapidly scales up investment in its own open source capabilities. Speaking at the 56th edition of the gathering in Switzerland, Schmidt drew a sharp contrast between the dominant approaches in the United States and China, and warned that Europe’s hesitation could leave it reliant on foreign technology at a critical moment.

In his remarks, Schmidt noted that most US companies are moving toward proprietary, closed source models that are licensed and controlled by a small number of firms. By contrast, he said, China is embracing open weight and open source strategies, making its models widely accessible. Without a major financial commitment to build and maintain competitive European AI models, Schmidt argued, the continent will inevitably end up relying on Chinese systems. Such an outcome, he suggested, would carry strategic risks for Europe’s digital sovereignty and long term competitiveness.

Schmidt’s warning came as world leaders met under the forum’s theme of a spirit of dialogue, attempting to navigate an international landscape reshaped by rapid technological change and deepening geopolitical fragmentation. In a separate session, Microsoft chief executive Satya Nadella amplified the sense of urgency, describing AI output as a new kind of commodity measured in tokens, the units of processing that power AI tasks. He argued that the cost of energy will largely determine which countries emerge as winners in the AI race, tying economic performance directly to access to affordable power.

Nadella stressed that future gross domestic product growth in any country will be closely linked to the energy costs associated with deploying AI at scale. He pointed out that energy prices in Europe remain among the highest in the world, a legacy of sanctions and supply disruptions following Russia’s invasion of Ukraine. Against that backdrop, he urged European policymakers to adopt a broader global perspective and not focus solely on questions of technological sovereignty, implying that competitiveness and integration into global AI value chains are equally vital.

The energy challenge facing Europe is substantial. Research by the OECD indicates that electricity intensive European industries, including the data centres needed to run AI systems, face persistent cost disadvantages. Demand from data centres alone is projected to climb from around 96 terawatt hours in 2024 to roughly 236 terawatt hours by 2035, underscoring the scale of investment required in energy infrastructure and efficiency if Europe is to host competitive AI infrastructure on its own soil.

A parallel Chinese perspective framed the Davos discussions in more cooperative terms. In a commentary published before the forum, CGTN analyst Belunn Se argued that cooperation rather than confrontation has become a structural requirement of global development, necessary to address challenges ranging from climate change to AI governance. The piece said China’s approach to AI regulation is shifting from a principles based orientation toward more institutionalised and systematic frameworks, positioning Beijing as a rule shaper in emerging global governance structures rather than a mere rule taker.

This year’s forum drew a record level of government participation, with more than 60 heads of state and government joining central bank leaders and ministers from over 130 countries. The Atlantic Council observed that the AI race in 2026 remains defined by a multipolar order in which the United States and China continue to exert the greatest influence, while middle powers, including European states and India, work to narrow the gap and carve out space for their own ecosystems.

Schmidt’s concerns tap into broader unease about the geopolitical implications of open source AI. He has previously argued that the vast majority of governments and countries without the financial resources of the West will end up adopting Chinese models as a default, not because they are technically superior but because they are available at low or no cost. Chinese models developed by firms such as DeepSeek and Alibaba have gained traction globally, and some US technology companies are already integrating them into applications, highlighting how quickly the balance of influence can shift.

European officials have begun to respond. The European Commission has launched consultations on strengthening the continent’s open source AI ecosystem, acknowledging that a large share of the value created by open source projects originating in Europe is currently captured elsewhere. For policymakers, investors and technology professionals, the debate in Davos underlined a critical choice for Europe: whether to invest aggressively in homegrown AI models and infrastructure, or risk seeing its digital future shaped by systems developed in Beijing and California.