Brian Letort, Head of Data office & AI, and Kadri Linask-Goode, Director, Privacy, Digital Realty

April 3, 2025

Businesses around the world are committed to the growth of artificial intelligence (AI) — and now countries are, too.

IDC predicts that by 2028, the spending on AI infrastructure globally will surpass $100 billion, and over the next five years, the compound annual growth rate of spending on AI infrastructure in Asia Pacific will grow by 20%, 16% in the US, and EMEA is expected to rise by 13%. One major area of investment for countries will be sovereign AI.

Since AI’s tipping point into widespread adoption with OpenAI's introduction of ChatGPT, businesses, organisations, and other entities have been creating their own AI models or training LLMs with their own data.

However, this does beg the question: if AI is to be open source, who controls that data, and are there instances where that data may conflict with a country's culture, economic goals, or national security?

This is why sovereign AI, where a nation creates or adopts its own proprietary AI, is increasingly becoming an appealing priority for nations. Let’s explore this emerging competitive landscape and how sovereign AI could shift access and use of AI into the future.

The growing importance of sovereign AI and the history of data sovereignty

There’s a global race to the top in AI today — or at least a race to compete on a global scale with other AI leaders.

As AI continues to grow, countries face the question of whether they want to rely on AI models developed in other countries or direct their own country's technological future.

Many are choosing the latter and are now looking at ways to develop AI built on their own data, infrastructures, and networks. Sovereign AI is a local approach that creates a more confined, controlled environment, especially when it comes to meeting stricter data privacy regulations. With sovereign AI, a country or region extends sovereignty into model training, deployment, and usage.

The idea of a region owning its AI isn't necessarily a new concept. Data sovereignty dictates that data should be governed by the rules and regulations in the region where it’s generated, collected, or stored. That idea extends to sovereign AI, but goes one step further.

Instead of simply governing the data in that region, AI is built on that country's unique data, research, intelligence, and history. Because the infrastructure is developed and housed in-country, sovereign AI models address security challenges and reduce the risk of relying on foreign AI models.

For example, considering the economic and political conditions within the European Union (EU) today — data protection becomes more critical. Sovereign AI could be a source of digital innovation that suits the fragmented data transformation in the EU. There are very different jurisdictional and political strategies across the EU and wider EMEA, and sovereign AI can give nations a competitive advantage in the era of digital transformation.

The key benefits of sovereign AI

The biggest benefit sovereign AI offers is that it increases data privacy and security domestically. Instead of relying on AI developed elsewhere, countries can control every aspect of their home-grown AI models. This includes what data is being used to train it to how it's being run and housed.

Keeping their AI model in-country also allows for increased control over security parameters and data privacy, and it can be tailored to national security needs as well. This is especially important when Gartner predicts that by 2027, 40% of data breaches will be due to misuse across borders.

Additionally, sovereign AI models can be tailored to meet the unique needs of the country with data that is contextual to the country's culture, history, and sociological nuances of that country. The AI model can be trained on that country's research and national contributions, like healthcare and technological advancements. It’s AI that “speaks the same language” as the country it’s in.

Gartner also predicts that by 2027, a robust approach to data governance will be a requirement of “sovereign AI laws and regulations worldwide.” Sovereign AI essentially necessitates good governance and forces the application of data localisation, where data is collected, processed, and stored in a specific region. This can allow for faster access and lower latency, making AI use more efficient and impactful.

The challenges in building and implementing sovereign AI

While there may be a lot of upside to sovereign AI, there are a few big challenges countries may face when building their own models. If a country wants to create its own AI model, it must have the ability to make the investments to support those AI ambitions.

This includes having infrastructure in place that supports AI's high energy needs, data centre solutions with high-performance compute, GPU clusters for the parallel processing required to accelerate AI training, and advanced network connectivity.

Another challenge is building and maintaining a skilled workforce in AI development and management to support sovereign AI ambitions. According to a new report, Artificial Intelligence Engineering roles are the fastest growing jobs in the UK and the Netherlands, and it’s one of the fastest growing jobs in over a dozen countries.

Developing sovereign AI innovations that are competitive with the rest of the world means investing in AI education and training programs to execute those goals.

Along with talent is the need for robust regulatory frameworks to govern the development and use of sovereign AI. This includes how data is collected and used, how AI models are trained, transparency into data privacy, and addressing the many ethical questions that have arisen around AI’s use. Countries also need to establish governance for oversight, leadership, management, and accountability.

Sovereign AI in action and government-backed initiatives

We’re seeing developments in the advancement of sovereign AI, including governments pledging investments in innovation and development of AI in their own countries, as well as partnering with private companies to advance AI initiatives.

The United States has announced its commitment to work with private companies, including OpenAI and Microsoft, to advance the Stargate Project. The goal is to invest $500 billion into AI development in the United States over the next four years, including building data centres across the country.
In October 2024, Denmark launched its Gefion Supercomputer, which will use “AI to accelerate innovation in many areas, ranging from quantum computing to drug discovery, to societal challenges such as the transition to green energy,” according to the Danish Centre for AI Innovation. Having access to the computing power needed for this supercomputer was one of the roadblocks to innovation, but through public/private partnerships, Denmark now has the capabilities to move forward with their AI objectives.
At the AI Action Summit in February 2025, French President Emmanuel Macron pledged €109 billion in private investments to AI advancement to AI advancement and training, which includes building data centres and infrastructure to support those advances. While Mistral, a French AI company, has been working on open-source LLMs and offering a GPT rival called Le Chat, it’s unclear whether France’s newly announced investments would go towards backing Mistral and other French private companies, or towards funding a new government-owned AI model.
Also announced at the AI Action Summit in Paris was the creation of InvestAI, which seeks to raise €200 billion for investments in AI that include building AI gigafactories across the EU. In a statement, the president of the European Commission Ursula von der Leyen said that “AI will improve our healthcare, spur our research and innovation and boost our competitiveness. We want AI to be a force for good and for growth. We are doing this through our own European approach.”
Italy, Germany, and Sweden have also announced strategic partnerships and technological developments that will increase AI and AI investments in their respective countries. The UAE, which also pledged investments in France’s AI initiatives, is looking to advance its AI initiatives through a dedicated AI research institute and with the development of its own LLM called Falcon.

The future of sovereign AI

Sovereign AI is relatively new and we’re just seeing the first investments in its development. As these initiatives grow, we’ll see if countries decide to create sovereign AI models entirely within their government, reach out to private companies to back independent research and development, or a hybrid of both.

IT leaders in countries investing more heavily in top-down AI initiatives need to be prepared not just for new sovereign AI models, but for what the impacts on data sovereignty will be in those regions as a whole.

What’s clear is that the enormous amount of effort and investments being put behind AI’s evolution will benefit AI advancements in both the public and private sectors for years to come.

Digital Realty’s PlatformDIGITAL® is the home of AI. We have the core understanding and expertise required for generative AI and high-performance compute to be successful.

Digital Realty helps scale AI deployments across environments with our global data centre footprint, PlatformDIGITAL®, and a Pervasive Datacenter Methodology® that's proven to optimise data exchange.

Learn more about how we enable data privacy, sovereignty, and private AI with the customer use cases we see most often and contact us today to schedule a workshop with our experts.

Looking to learn how to successfully implement AI to bring your organisation into the future? Download the “AI for IT Leaders: Deploying a Future-Proof IT Infrastructure” today.

XTREME-D Inc. Rapidly Scales High-Performance Computing Services to Support New Clients on PlatformDIGITAL®