The quietly revolutionary announcements in the EU’s data union strategy

The breakneck growth of AI companies these past three years has made it apparent that Europe is not the biggest fish in the technological pond. 

Compared to the eye-watering valuations of American AI companies and the mammoth infrastructure projects announced to underpin them, incremental policy shifts within Europe have largely gone unnoticed.

In November, the EU released its data union strategy and a strategic roadmap for digitalisation and AI in energy is imminent. Europe will struggle to rival the deep pockets of the US and China, but these policies will unlock something more fundamental for Europe’s future competitiveness: data.

Europe’s paradox: advanced energy systems, weak AI scale

Europe is unlikely to win the race for general purpose AI. Large language models like ChatGPT have opened up a chasmic technological and financial lead. Where Europe can compete is on domain specific models because of its digitalised infrastructure. Energy is a prime example.

On many measures, Europe’s energy sector is more digitally advanced than the US or China. Renewable penetration is approaching 50%, electricity markets are liberalised, smart meters are near universal in a number of countries, and electric vehicle and heat pump adoption is accelerating. The foundations of a digital energy system are already in place.

Yet despite these advantages, AI uptake in energy remains low. A recent survey of energy utilities shows that, while 91% have experimented with or implemented generative AI, nearly half are still in the early stages of adoption, and only 7% report advanced use. The biggest barrier to AI deployment? Access to high-quality data.

Data as a strategic advantage

Data is foundational for AI development. In energy, vast quantities of data are already collected, but privacy, competition and cybersecurity rules often limit access. The result has been a proliferation of initiatives, such as the common European energy data space, that have struggled to translate ambition into reality – fast.

The data union strategy signals a course correction. Rather than doubling down on building more infrastructure, the Commission is shifting focus towards frameworks that improve accessibility. Clarity for data pooling and championing of data quality standards will make it easier for companies to share data with confidence. But perhaps the most significant signal is support for synthetic data.

From data scarcity to data abundance

Synthetic data is artificially generated data that’s statistically representative of real data, without linking to real individuals or assets. Already used in healthcare and finance to preserve privacy and overcome data scarcity, it enables large-scale model training, with real data used for fine-tuning and validation.

This shift is already gaining traction, not least evident by Nvidia’s acquisition of Gretel.ai, reflecting growing recognition that data quality, not just compute, shapes model performance.

Access to large volumes of data enables the development of better targeted models. Synthetic demand data, can improve load forecasting and investment planning; synthetic generation data can help models anticipate output under extreme weather; and synthetic outage data can strengthen fault detection and response. 

Models grounded in reality are more likely to be trusted and adopted, creating a virtuous circle between model quality, deployment and impact.

This is critical for Europe’s energy security and competitiveness. Making synthetic data a strategic enabler of European AI leadership supports both sides of the AI equation: developing higher quality, domain specific models, which can then be deployed across real energy systems at scale.

Have you read:
EU Energy Projects Podcast: How the EU plans to support AI in the energy sector
Europe’s data centres industry on brink of unprecedented expansion

Building on open source initiatives

Rather than starting from scratch, the EU should build on open source initiatives that already exist across the region. Open ecosystems tend to move faster and adapt more easily than consortia designed to satisfy funding calls. Just as openly available datasets catalysed innovation in other sectors, open initiatives can do the same for energy.

Efforts such as OpenSynth, our collaboration with Linux Foundation Energy, which focuses on sharing synthetic data, alongside initiatives like GridFM, which is developing a foundation model for the electricity grid using synthetic data, show how shared data, models and standards can co-evolve. 

Their experience could directly inform the Commission’s proposed guidance, standards and certification schemes for synthetic data. Policies that legitimise and scale open approaches would help turn Europe’s data advantage into deployable AI at speed.

Strategic roadmap

As we await the publication of the strategic roadmap for digitalisation and AI in energy, it is yet to be seen whether Europe will move quickly enough from policy to implementation.

Synthetic data factories must become enablers, not bottlenecks. Standards should accelerate, rather than delay, adoption. Open source, cross-border collaboration should be treated as a competitive strength, not a risk. 

Europe can successfully develop and deploy AI in energy by turning data into a strategic asset. If it is to marry its clean power ambitions with AI progress, it will need to make the most of its inherent advantages – such as its digitised energy system. 

There is a long way to go, but the data union strategy is a quiet and significant step in that direction.

About the author