The AI race is creating an unusual global split. Countries hungry for technology investment are courting data centers as strategic infrastructure at the same moment communities in parts of the United States and mature data center markets are questioning how much more they want. Electricity prices, water use, grid construction and land have pushed what once was a low-profile infrastructure business into public debate.

That leaves a cloud hanging over the AI boom. Can countries capture the economic and strategic value of new computing infrastructure without producing the resistance now complicating construction elsewhere?

Lithuania offers an interesting early case. Its government has made AI infrastructure part of its national strategy, is backing a €130 million national AI factory and is actively seeking private data center investment. Yet Lithuania has not experienced hyperscale development on the level seen in places such as Northern Virginia or Ireland. Its enthusiasm is real, but it has not yet been tested by the scale of growth that turned data centers into a political issue elsewhere.

Lithuania Wants Compute, Not Just AI Models

Lithuania’s recently approved AI guidelines put its goals in writing. One objective calls for “modern, secure, and technologically sovereign AI technology infrastructure.” The plan includes national cloud resources, national high performance computing capacity and better access to international computing resources. Another objective calls for attracting foreign direct investment into data and technology infrastructure.

The government approved the 2026 to 2035 guidelines in April. Their broader vision ties AI to economic value, public administration and the preservation of sovereignty.

Economy and Innovation Minister Edvinas Grikšas, speaking at the Startup Fair in Vilnius and in a direct one-on-one interview,said that Lithuania is approaching sovereignty “more from the infrastructure side.” He framed European technological capacity as something that can grow without shutting out companies from allied countries.

That distinction gives Lithuania room to court American technology companies at the same time it seeks greater European control over computing infrastructure.

Diana Girdenytė, a strategist in Invest Lithuania’s Strategy and Product Development Department, shared in an interview that “open source models are particularly relevant from an investment attraction perspective, as running them requires physical computing infrastructure, which is exactly the type of infrastructure we are looking to attract to Lithuania,” she said in an interview.

A country does not necessarily need to create a frontier model that competes directly with the largest American or Chinese systems. It can build computing capacity at home, create national cloud resources, provide access to high performance computing and give local companies more choices about where their models and data run.

Lithuania is not limiting its search to companies that build and operate data centers. It wants big end users customers too, including large technology companies that can fill those facilities with computing equipment.

Data Centers Have Become A Community Issue

The mood for data center growth looks very different in parts of the United States and Europe.

A Reuters and Ipsos poll cited in the report found only 14% of Americans surveyed said they would welcome a data center in their own community. On July 18, opponents staged 142 coordinated data center protests in 42 states. Residents raised questions about water, power demand, environmental effects, consultation with communities and whether the facilities create enough local economic benefit to justify their resource use.

Electricity is one reason the fight has intensified. Lawrence Berkeley National Laboratory estimates that data centers could account for 11.8% of total U.S. electricity use in 2030 under its reference case. Its modeled range runs from 9.5% to 15.3%. Importantly, those are national figures. The burden is often much more concentrated in regions where large clusters of facilities connect to the same grid.

The issue has moved into Congress. On September 16, the U.S. House passed the Ratepayer Protection Act by 417 votes to 3. The measure directs state utility regulators to examine whether very large electricity users such as data centers should carry the added infrastructure costs created by their demand rather than passing those costs to households. Critics cited by Reuters said the bill does not go far enough, since it requires consideration of such protections rather than imposing a single national cost allocation rule.

Ireland Shows What Happens When Growth Hits The Grid

Ireland has become one of Europe’s major data center markets, helped by its technology industry and role as a European base for U.S. companies. The success comes with a heavy electricity requirement.

Ireland’s Central Statistics Office reported that data centers consumed 23% of the country’s metered electricity in 2025, up from 5% in 2015. Their electricity consumption reached 7,663 gigawatt hours in 2025, a 10% increase in just one year.

While Ireland has not closed the door to new facilities, it has changed the conditions.In December 2025, Ireland’s Commission for Regulation of Utilities adopted a new electricity connection policy requiring new data centers seeking grid connections to provide generation or storage capacity, either on site or nearby, that matches their requested maximum import capacity. The regulator said the rules were designed to create a route for future connections that accounts for grid constraints, generation adequacy and renewable energy goals.

Singapore is taking a controlled expansion route. Its government opened at least 200 megawatts of new data center capacity under an allocation program announced in December 2025. Applicants face demanding efficiency standards, including a power usage effectiveness target of 1.25 or better and a requirement that at least half of the new capacity use eligible green energy pathways.

Lithuania is trying to get ahead of these power demands and constraints while also balancing and factoring in community needs and demands.

“We are more than open, actually,” Grikšas said when I asked about the community resistance appearing in the United States.

Lithuania’s Energy Ministry says its Kruonis Technology Park can provide about 500 megawatts of electrical capacity for data center demand in the near term, with longer term capacity that could exceed one gigawatt. The site is beside the 900 megawatt Kruonis pumped storage plant and connected to three independent 330 kilovolt transmission lines.

The same ministry says roughly 76% of electricity generated in Lithuania comes from renewable sources. Lithuania aims for renewable generation to cover all final electricity consumption by 2028. That figure is a national target, not a promise that a future hyperscale campus would arrive without new grid or generation requirements. A very large data center would itself change the demand equation.

Girdenytė said Lithuania does not yet have a hyperscale data center and that government bodies are aligned around attracting at least one or two. She said officials had not encountered the type of public opposition visible in some U.S. communities, noting that sites being considered are outside dense residential areas. Of course, that is the view from officials working to attract the projects, rather than evidence that future projects would face no local objections.

Sovereign AI Is Becoming A Physical Infrastructure Question

For years, much of the sovereignty debate centered on models, data and regulation. Much of the conversation applies to whose models are used and whether or not countries and companies exert control over their usage. But many other questions apply to the infrastructure tier. Where are the GPUs? Who supplies their electricity? Who controls the cloud capacity? Can a European startup get enough compute without depending entirely on infrastructure outside the region?

There is a second set of questions that is not as easily settled in an AI strategy document. Who pays for the grid upgrades? How much power can one project absorb before other users feel the effects? What happens when a welcoming national government encounters a less enthusiastic municipality?

Lithuania launched a €130 million LitAI project that puts the government directly into the compute business. With half of the project budget due to come from Lithuania and half from European Union programs through the EuroHPC Joint Undertaking and with Vilnius University leading the consortium, EuroHPC describes LitAI as an effort to turn Lithuania’s high performance computing capacity into “sovereign AI optimised infrastructure” serving national and European needs.

Countries like Lithuania have an advantage of arriving at their stage later than some major data center markets. It can see the fights over electricity costs in the United States, the grid rules adopted in Ireland and the constrained allocation model used in Singapore before hyperscale development takes hold at home. The hope is that the demand can be met with local supply. For now, Lithuania is rolling out the welcome mat.