News
Lessons from the Nebius deal: how can Estonia attract the next wave of AI investment?
Lessons from the Nebius deal: how can Estonia attract the next wave of AI investment?
Nebius’s 22 MW AI factory at the Greenergy data centre in Hüüru is set to become operational later this year. The road to this milestone was far from straightforward, making it worth reflecting on the lessons learned so that this landmark deal can pave the way for further investment, write RASK attorneys-at-law Karl Erik Esko and Karl-Kristjan Kahm.
The first step for the representatives of Greenergy and Tensor Estate was to explain to international AI infrastructure developers what and where Estonia is, and what opportunities the country has to offer. It was equally important to address the potential risks and, above all, explain how they could be mitigated. Ultimately, speed became one of the decisive factors, along with the fact that Greenergy’s existing facility made it possible to offer a turnkey solution within a short timeframe. This proved to be a compelling advantage.
As a pioneering transaction, the deal therefore provides valuable insight into where Estonia should focus its efforts when competing for future investment, as well as the potential bottlenecks that could stand in the way.
The Dutch company Nebius provides businesses with computing capacity and cloud services for training and running artificial intelligence models. Microsoft, Amazon, Google and Meta are among the biggest drivers of the global wave of investment in AI infrastructure. Alongside these technology giants, a new category of cloud providers specialising specifically in AI – so-called neocloud companies such as Nebius and CoreWeave – has rapidly emerged.
For Estonia, this means that future investment should be sought from both groups. A small domestic market is not necessarily a constraint, as a data centre located in Estonia can serve customers around the world. The pace of investment also illustrates the scale of the market. According to an analysis published by the International Energy Agency in April, the capital expenditure of five major technology companies exceeded USD 400 billion in 2025 and was projected to increase by a further 75% in 2026. A significant share of this expenditure is related to data centres. Google’s announcement in September that it plans to invest at least EUR 13 billion in Finland’s digital infrastructure in 2027–2028 demonstrates that investment on this scale is also reaching Estonia’s immediate region.
The clearest lesson from our experience with the transaction is simple: electricity comes first and time second. The availability of the former is a prerequisite for entering negotiations; the latter is critical to bringing them to a successful conclusion. A project that looks excellent on paper will remain an Excel spreadsheet if its potential completion date does not align with the customer’s plans. And in this market, time is measured in weeks rather than months.
So what can Estonia compete on against its neighbours? Although electricity in Estonia is relatively inexpensive in the European context, it is not our greatest advantage compared with neighbouring countries. In 2025, the average wholesale electricity price was approximately EUR 80 per MWh in Estonia and EUR 40 per MWh in Finland. While these figures do not represent the final cost of electricity for a data centre, the difference is significant. The countries are also in very different weight classes in terms of the size of their electricity systems: Estonia reached its all-time peak consumption of approximately 1,723 MW in February 2026, while Finland’s record reached 15,600 MW in January of the same year. Even taking into account Finland’s population, which is almost five times larger, it is difficult to establish a direct relationship between consumption levels and electricity prices; the factors determining electricity prices lie elsewhere. Estonia should therefore look for its competitive advantage elsewhere too, particularly in the feasibility of individual projects and the speed of decision-making.
A favourable climate and a functioning rule of law are unlikely, on their own, to convince major technology companies, as other Nordic countries offer the same advantages. Estonia has an opportunity to distinguish itself as a fast and reliable partner – and to demonstrate this in practice. In addition to a capable private sector and strong international networks, this requires competent and efficient public authorities and procedures that facilitate cooperation and the identification of solutions. Improving these processes and promoting Estonia’s capabilities internationally is an area in which the state should invest deliberately.
Estonia could also differentiate itself through its ability to deliver the necessary grid connections within a reasonable timeframe. In European data centre development, alongside the cost of land and electricity, an increasingly critical factor is the point at which the required power capacity will actually become available. In parts of Europe, developers may have to wait years for a large-scale grid connection. A project can therefore lose its competitive position on schedule alone. In Estonia, by contrast, it is possible to realistically model both the completion timeline and the financial viability of a data centre when making the investment decision.
This, in turn, demonstrates how closely the potential of data centres is linked to the development of the energy system as a whole. Investment in generation, storage, networks and consumption must progress in a sufficiently coordinated manner to ensure that one does not become a constraint on the development of another. The state’s primary role is to create as clear and predictable an environment as possible for electricity grid development, thereby reducing the uncertainty associated with long-term investment.
Finland also offers a useful example in this respect. According to Fortum, its long-term power purchase agreement with Google will support investments required to extend the operating life of the Loviisa nuclear power plant. Estonia should likewise look at data centre development in broader terms than simply securing individual investments and explore ways of linking their growth to new energy generation and the development of local businesses.
One of the advantages of a small country should be its ability to bring developers, customers, grid operators, local authorities and government agencies around the same table quickly. The fast-track procedure for strategic investments currently under consideration by the Estonian Parliament could become a practical tool for facilitating this kind of cooperation. Speed, however, is valuable only if the quality and predictability of decision-making are maintained.
Estonia’s speed becomes a genuine competitive advantage only if investors can also rely on projects being delivered in practice. The role of legal counsel is to translate commitments regarding power capacity and completion dates into the terms governing the investment, construction and lease arrangements. Among other things, this means agreeing on who bears the risks associated with delays and unforeseen costs, and how the project can move forward if difficulties arise. A sensible allocation of risk also takes into account which party is actually in a position to control or influence a particular risk. It is precisely these arrangements that give investors, developers and customers the confidence to undertake substantial long-term commitments before a data centre has even been completed.
The first deal has now opened the door, and its successful launch could open it further for future investment. It would be unrealistic to expect a single investment to create an entire AI economy in Estonia overnight. However, if Estonia can demonstrate its capabilities, it will be in a stronger position to discuss the broader benefits as well: access to computing capacity for Estonian businesses and researchers, investment in energy infrastructure, smart energy solutions, and maximising local economic impact and job creation. These objectives should be discussed as early as the planning stage of a new project, while keeping Estonia’s expectations and messaging clear.
The rapid development of AI infrastructure gives Estonia an opportunity to participate in this technological transformation, but attracting the next data centre alone will not be enough. Greater value is created when new investment also drives the development of local businesses and skills, new energy generation and access to computing capacity. Estonia’s opportunity lies in demonstrating that it can provide the necessary capacity at the right time and turn individual investments into broader economic impact.
* RASK advised Tensor Estate on its investment in the Greenergy data centre and advised Tensor Estate and the Greenergy data centre on the lease agreement with Nebius. For the latter transaction, the team also worked alongside Simmons & Simmons (UK), one of Europe’s leading law firms for AI factory projects.
The first step for the representatives of Greenergy and Tensor Estate was to explain to international AI infrastructure developers what and where Estonia is, and what opportunities the country has to offer. It was equally important to address the potential risks and, above all, explain how they could be mitigated. Ultimately, speed became one of the decisive factors, along with the fact that Greenergy’s existing facility made it possible to offer a turnkey solution within a short timeframe. This proved to be a compelling advantage.
As a pioneering transaction, the deal therefore provides valuable insight into where Estonia should focus its efforts when competing for future investment, as well as the potential bottlenecks that could stand in the way.
The Dutch company Nebius provides businesses with computing capacity and cloud services for training and running artificial intelligence models. Microsoft, Amazon, Google and Meta are among the biggest drivers of the global wave of investment in AI infrastructure. Alongside these technology giants, a new category of cloud providers specialising specifically in AI – so-called neocloud companies such as Nebius and CoreWeave – has rapidly emerged.
For Estonia, this means that future investment should be sought from both groups. A small domestic market is not necessarily a constraint, as a data centre located in Estonia can serve customers around the world. The pace of investment also illustrates the scale of the market. According to an analysis published by the International Energy Agency in April, the capital expenditure of five major technology companies exceeded USD 400 billion in 2025 and was projected to increase by a further 75% in 2026. A significant share of this expenditure is related to data centres. Google’s announcement in September that it plans to invest at least EUR 13 billion in Finland’s digital infrastructure in 2027–2028 demonstrates that investment on this scale is also reaching Estonia’s immediate region.
Cheap electricity vs speed of decision-making
The clearest lesson from our experience with the transaction is simple: electricity comes first and time second. The availability of the former is a prerequisite for entering negotiations; the latter is critical to bringing them to a successful conclusion. A project that looks excellent on paper will remain an Excel spreadsheet if its potential completion date does not align with the customer’s plans. And in this market, time is measured in weeks rather than months.
So what can Estonia compete on against its neighbours? Although electricity in Estonia is relatively inexpensive in the European context, it is not our greatest advantage compared with neighbouring countries. In 2025, the average wholesale electricity price was approximately EUR 80 per MWh in Estonia and EUR 40 per MWh in Finland. While these figures do not represent the final cost of electricity for a data centre, the difference is significant. The countries are also in very different weight classes in terms of the size of their electricity systems: Estonia reached its all-time peak consumption of approximately 1,723 MW in February 2026, while Finland’s record reached 15,600 MW in January of the same year. Even taking into account Finland’s population, which is almost five times larger, it is difficult to establish a direct relationship between consumption levels and electricity prices; the factors determining electricity prices lie elsewhere. Estonia should therefore look for its competitive advantage elsewhere too, particularly in the feasibility of individual projects and the speed of decision-making.
A favourable climate and a functioning rule of law are unlikely, on their own, to convince major technology companies, as other Nordic countries offer the same advantages. Estonia has an opportunity to distinguish itself as a fast and reliable partner – and to demonstrate this in practice. In addition to a capable private sector and strong international networks, this requires competent and efficient public authorities and procedures that facilitate cooperation and the identification of solutions. Improving these processes and promoting Estonia’s capabilities internationally is an area in which the state should invest deliberately.
Estonia could also differentiate itself through its ability to deliver the necessary grid connections within a reasonable timeframe. In European data centre development, alongside the cost of land and electricity, an increasingly critical factor is the point at which the required power capacity will actually become available. In parts of Europe, developers may have to wait years for a large-scale grid connection. A project can therefore lose its competitive position on schedule alone. In Estonia, by contrast, it is possible to realistically model both the completion timeline and the financial viability of a data centre when making the investment decision.
This, in turn, demonstrates how closely the potential of data centres is linked to the development of the energy system as a whole. Investment in generation, storage, networks and consumption must progress in a sufficiently coordinated manner to ensure that one does not become a constraint on the development of another. The state’s primary role is to create as clear and predictable an environment as possible for electricity grid development, thereby reducing the uncertainty associated with long-term investment.
Finland also offers a useful example in this respect. According to Fortum, its long-term power purchase agreement with Google will support investments required to extend the operating life of the Loviisa nuclear power plant. Estonia should likewise look at data centre development in broader terms than simply securing individual investments and explore ways of linking their growth to new energy generation and the development of local businesses.
Investors need certainty, not just promises
One of the advantages of a small country should be its ability to bring developers, customers, grid operators, local authorities and government agencies around the same table quickly. The fast-track procedure for strategic investments currently under consideration by the Estonian Parliament could become a practical tool for facilitating this kind of cooperation. Speed, however, is valuable only if the quality and predictability of decision-making are maintained.
Estonia’s speed becomes a genuine competitive advantage only if investors can also rely on projects being delivered in practice. The role of legal counsel is to translate commitments regarding power capacity and completion dates into the terms governing the investment, construction and lease arrangements. Among other things, this means agreeing on who bears the risks associated with delays and unforeseen costs, and how the project can move forward if difficulties arise. A sensible allocation of risk also takes into account which party is actually in a position to control or influence a particular risk. It is precisely these arrangements that give investors, developers and customers the confidence to undertake substantial long-term commitments before a data centre has even been completed.
The first deal has now opened the door, and its successful launch could open it further for future investment. It would be unrealistic to expect a single investment to create an entire AI economy in Estonia overnight. However, if Estonia can demonstrate its capabilities, it will be in a stronger position to discuss the broader benefits as well: access to computing capacity for Estonian businesses and researchers, investment in energy infrastructure, smart energy solutions, and maximising local economic impact and job creation. These objectives should be discussed as early as the planning stage of a new project, while keeping Estonia’s expectations and messaging clear.
The rapid development of AI infrastructure gives Estonia an opportunity to participate in this technological transformation, but attracting the next data centre alone will not be enough. Greater value is created when new investment also drives the development of local businesses and skills, new energy generation and access to computing capacity. Estonia’s opportunity lies in demonstrating that it can provide the necessary capacity at the right time and turn individual investments into broader economic impact.
* RASK advised Tensor Estate on its investment in the Greenergy data centre and advised Tensor Estate and the Greenergy data centre on the lease agreement with Nebius. For the latter transaction, the team also worked alongside Simmons & Simmons (UK), one of Europe’s leading law firms for AI factory projects.