Does Poland have a chance to become a leader on the ‘AI Continent’?

A few weeks ago, I received a document from the Ministry of Digital Affairs for review with the telling title ‘Policy for the development of artificial intelligence in Poland until 2030’. It is still too early for me to discuss its contents in detail (after all, consultations are still ongoing). But the plans are very ambitious. I don’t want to say that they are overestimated, because in an efficient country they would be absolutely achievable.

And that’s the rub. We don’t live in an efficient country. And I said ‘I’ll take you up on that’.

My benchmark is France, which is the undisputed leader in Europe when it comes to creating large language models (LLMs). We are not going to compare ourselves to the US or China, because it makes no sense. Let’s stay on the continent and have the same goals as other medium-sized countries in the region. And let’s focus on just one, but key aspect – energy, without which we can only dream of ‘AI factories’.

By way of introduction – a few words about the challenges

Poland’s ambitions fit quite well into the global trend, and this should come as no surprise. In short, whoever masters artificial intelligence will win the technological race and, consequently, the economic and political race in today’s world. However, realising these aspirations requires not only advanced research, human capital development and significant technological investments, but above all, securing huge and, crucially, stable energy resources. The infrastructure necessary for the development of AI, in particular data centres processing huge amounts of information and supercomputers used to train LLMs, is characterised by exceptionally high electricity demand. The current state of the Polish energy sector, which (despite ongoing transformation) still relies heavily on coal, poses a fundamental challenge (if not a problem) for the sustainable and competitive development of AI in the country. I am convinced that our leaders are aware of this challenge, but prefer to ignore it.

Let me remind those who do not want to remember and tell those who do not yet know. Ensuring synergy between the AI development strategy and a coherent, long-term energy policy is essential, and one cannot succeed without the other. In this context, it is worth looking at the experience of France, which is currently the only country in Europe creating global LLM models and actively integrating AI development into its energy strategy, based largely on low-carbon and stable nuclear and renewable sources.

Polish strategic documents, such as PEP2040 and KPEiK, must therefore be updated to include this issue. And then these updates must be implemented, not just left on paper, if we want to succeed with AI. On the one hand, the need to meet AI’s needs may force faster investment in modern, low-carbon energy sources and in smart grids capable of managing new, large loads. On the other hand, the lack of strategic planning and integration of these two areas carries serious risks. A sharp increase in demand from AI, if not met by clean and stable sources, could paradoxically deepen dependence on fossil fuels, for example through the need to maintain or build new coal or gas capacity. This, leaving aside EU policy and climate targets, will force us to buy these fuels from Russia or the US. The French model, where low-carbon nuclear energy is seen as an asset for AI development, provides valuable guidance here.

Poland’s long-term competitiveness as a location for AI development is a case of killing at least two birds with one stone. The second (and perhaps even the first, if we do not ultimately build these ‘AI factories’) is attracting investors. Investment incentives for global players may not be enough for those investors and operators who attach importance to ESG and decarbonisation. And there are many of them. By offering energy with a high carbon footprint and potentially higher costs (related to CO2 emission fees and outdated infrastructure), Poland may lose the competition for key AI investments to countries such as France, which promote their decarbonised and stable energy sector as a strategic asset.

Subscribe so you don’t miss any new articles!

Some figures – global and national perspectives

According to the International Energy Agency (IEA), energy consumption by data centres worldwide, which was estimated at 460 TWh in 2024, and according to forecasts, we can expect this consumption to increase from 945 to over 1,000 TWh by 2030, which is roughly equivalent to Japan’s total annual electricity consumption.

The process of training large language models is particularly energy-intensive. For example, it has been estimated that training the GPT-3 model consumed nearly 1,300 MWh of energy, which is comparable to the annual energy consumption of approximately 130 American households. Analyses indicate that the computing power required to train increasingly advanced AI models is growing exponentially, roughly doubling every 3-4 months since 2012. What’s more, not only training, but also everyday interactions with AI systems, such as popular chatbots, can consume significantly more energy than standard web search queries. The scale of this increase is so significant that the IEA warns that the combined increase in energy consumption by data centres, cryptocurrency technologies and AI between 2022 and 2026 could be equivalent to the total annual electricity consumption of countries such as Sweden or Germany. In Europe, AI’s energy demand is expected to account for 4% to 5% of total electricity consumption by 2030, up from 2-3% in 2024.

In Poland, we are also seeing dynamic growth in energy demand from the data centre sector. According to data from the Ministry of Digital Affairs, in November 2024, total energy consumption by data centres in Poland reached 1.7 TWh, an increase of 23% year-on-year. PMR’s market forecasts indicate that the capacity of the Polish data centre market could almost triple between 2024 and 2030, from 173 MW to over 500 MW.

Don’t get me wrong, not everything in our country is painted in black. Polskie Sieci Elektroenergetyczne (PSE), the national transmission system operator, recognises this trend and has taken the increase in demand from data centres into account in its infrastructure development plans, reserving 1,200 MW of capacity for the development of these facilities by 2034. The implementation of specialised power supply systems dedicated to AI-generated loads is also being observed, and the power dedicated to artificial intelligence in Polish data centres increased by an impressive 94.3% in August 2024. However, it must be said that this is only a step in the right direction and we certainly cannot stop here if we want to think about becoming a leader.

And here we come to the heart of the matter:

This surge in energy demand from AI is putting enormous pressure not only on the development of new generation capacity, but also, and perhaps above all, on the modernisation and expansion of transmission and distribution networks. Data centres dedicated to artificial intelligence have specific load profiles – high power density and potentially large and rapid fluctuations in demand, which requires the power grid to be much more flexible, stable and responsive. Simply increasing the amount of energy available will not solve the problem if the grid is unable to reliably deliver and handle these dynamic, often unpredictable loads.

What is the current state of the energy sector in Poland?

An analysis of the current state of the Polish energy sector is crucial to assessing its ability to support Poland’s ambitions as an AI leader. This applies to the structure of energy production, emissions, the state of the network infrastructure and the political and strategic framework.

Energy mix and energy production

Poland’s energy sector is, of course, based on coal, but a gradual transformation is underway. In April 2025, we reached a milestone as the share of coal fell below 50% (49.4%), but in 2024 as a whole, coal accounted for 56.7% of energy production. In 2015, this figure was 87%, which shows the scale of the change and the challenges ahead.

The importance of RES is growing steadily. In April 2025, RES supplied 34.2% of energy, and in 2024 as a whole, 29-30%. The main drivers of growth are photovoltaics (over 22 GW of capacity in April 2025) and wind energy (10.8 GW). The role of natural gas as a transition fuel is also (unfortunately) growing. Energy production from gas increased by 44.2% (April 2025 y/y), which not only raises energy production costs but also makes us dependent on gas imports from the US and Russia. The total installed capacity in the National Power System at the end of 2024 was 72.2 GW (40.3 GW conventional, 31.8 GW RES). Despite progress, the Polish energy sector has one of the highest emissions in the EU (690-724 gCO2/kWh), significantly higher than in France (39-55 gCO2/kWh). The emissions of the French energy sector are therefore between 5 and 7% compared to Poland.

Network infrastructure

The state of the Polish transmission and distribution network is a serious challenge for the energy transition and the development of energy-intensive sectors such as AI, which makes it difficult for Poland to become a leader in AI. The network must cope with the growing share of unstable RES and the connection of new, large consumers.

The power system in Poland is struggling with balancing problems. In the second half of 2024, the imbalance exceeded 1,000 MW in 20% of billing periods. There are regular surpluses of energy from PV and power shortages at peak demand, which threatens the stability of the network. Connection restrictions and curtailment of RES are another problem. Between 2015 and 2021, nearly 6,000 connection refusals were issued (approx. 30 GW of potential). It is estimated that in 2030, RES generation restrictions may affect 11%, and in 2040 even 40%. In my opinion, this is not a desirable direction.

PSE is planning modernisation and development investments (PRSP for 2025-2034 worth over PLN 64 billion), including the construction of new 400 kV lines and power stations. This is to enable the connection of power from offshore wind farms, PV and onshore wind farms. Experts point to the need for proactive network development planning that anticipates market needs, which is crucial for the ambition to build ‘AI factories’. The implementation of smart grids is progressing, but Poland lags behind European leaders (approx. 40% implementation of smart meters compared to 90-94% in France), which limits our potential in this area.

Energy policy and development strategies

The key documents of Polish energy policy are PEP2040 (from 2021) and KPEiK. PEP2040 assumed, among other things, a maximum 56% share of coal by 2030 (which has already been achieved), a minimum of 23% RES in final consumption (32% in production) by 2030, and the launch of the first nuclear power plant in 2033.

However, these documents can be considered outdated. Work on updating PEP2040 and KPEiK is delayed. The draft update of the KPEiK presents more ambitious targets: 56% of energy production from RES by 2030 and a reduction in the share of coal to 22%. A strategic element of the transition is the development of nuclear energy (the first power plant by 2033, with a target of 6-9 GW by 2043). The PRSP PSE takes into account the connection of nuclear power plants and SMRs. Energy storage and hydrogen are becoming increasingly important (Polish hydrogen strategy of 2021). In October 2024, the storage capacity of data centres reached 145 MWh (year-on-year increase).

Poland is also developing an AI strategy (‘AI Development Policy in Poland from 2020’). The government has declared its support, e.g. for the AI Factory project in Poznań. Global technology players are investing in cloud infrastructure and AI in Poland. However, there is a discrepancy between outdated energy strategies and the dynamic development of RES and AI needs. This ‘strategic paralysis’ creates uncertainty for investors, hampering long-term planning that is crucial to powering the digital transformation and our ambitions. Network stability issues and RES connection constraints are a barrier to ‘green AI’ in Poland. Failure to address these issues effectively may force the continuation of production from conventional sources, undermining climate targets and deterring global companies developing LLM models and manufacturing systems based on them. The costs of balancing an unstable grid will be passed on to consumers, including AI data centres, negatively affecting their competitiveness.

The French development model should be a lesson for Poland

We need to expand our knowledge of the French energy mix

In 2024, nuclear energy accounted for the majority of electricity production, with various sources citing figures ranging from 67% of total generation to 77.7% (for the EDF Group’s global operations) or even 86.6% (for EDF’s operations in France). As we have already said, this translates into very low emissions. In 2024, as much as 94% of the energy produced by EDF was decarbonised, and on a national scale, 95% of electricity generation came from CO2-free sources.

France is also consistently developing renewable energy sources. In 2024, RES (including a significant share of hydropower) accounted for between 12.4% (EDF in France) and 16.6% (EDF Group globally) of the energy mix, and some data indicate that RES (including hydropower) accounted for 28% of total electricity production. France’s strategic goal is to achieve a 40% share of RES in electricity production by 2030 and a 33% share of RES in gross final energy consumption by 2030. Installed photovoltaic capacity is growing particularly rapidly, reaching 23.7 GW in September 2024.

The total installed capacity in the French power system was approximately 135 GW in 2020 (in Poland, it was 49 GW during the same period). At the end of 2024, installed capacity in nuclear power was estimated at around 61.4 GW, in hydropower at 25.7 GW, in wind power (onshore and offshore) at 24.3 GW, and in photovoltaics also at 24.3 GW. In mid-2024, the total installed capacity in the system reached 149.3 GW.

France is rapidly modernising its networks and supporting high-demand sectors

France is one of the European leaders in the implementation of advanced network technologies. The Linky programme (which we are also slowly implementing in our country), carried out by the main distribution system operator Enedis (part of the EDF group), has led to the installation of 35.7 million smart electricity meters between 2015 and 2021. These devices not only enable remote reading of energy consumption and real-time monitoring by consumers, but also offer control functions for end devices such as electric vehicle chargers and electric water heaters. This significantly increases the flexibility of the power system and facilitates demand management.

The French are also actively using their energy advantages to attract investment in energy-intensive AI infrastructure, such as data centres (which we unfortunately still cannot offer). The country promotes the availability of decarbonised, abundant and stable electricity, a systematically expanded high-voltage network, the availability of suitable land for investment and simplified administrative procedures. The stability of the French electricity grid, based largely on predictable generation from nuclear power plants, is seen as a key factor for data centre operators, for whom power reliability is a priority. Data centres in France consumed approximately 8.5 TWh of electricity per year, accounting for approximately 2% of total national consumption.

Government strategies and investments in AI

Since 2017, France has been implementing (I emphasise this because they have not only created it, but are actually implementing it) a national strategy for the development of artificial intelligence, aiming to consolidate its position as a global leader in this field. As part of this strategy, investment plans worth over EUR 109 billion have been announced, to be allocated to the development of infrastructure projects supporting AI. These investments come from both public funds and private investors such as Brookfield and Iliad.

France is also actively involved in EU initiatives aimed at strengthening Europe’s AI capabilities. One example is the InvestAI programme, under which the European Union plans to mobilise €200 billion for investments in artificial intelligence, including €20 billion for the creation of European ‘AI gigafactories’ – specialised computing centres designed to train the most complex AI models. France also supports projects such as OpenEuroLLM, which aims to develop European multilingual language models. We would also like to have such models and factories, as I wrote earlier, but… Well.

French energy policy, as set out in documents such as the Multiannual Energy Programme (PPE) and the National Low-Carbon Strategy (SNBC) , is closely linked to the objectives of the digital transformation. The French government consistently emphasises the importance of stable and competitive energy prices and high network reliability for attracting and developing the AI sector. There are plans to further develop nuclear energy, including the construction of six new reactors, and to continue investing in renewable energy sources. In addition, France is pursuing an active ‘Green IT’ policy aimed at reducing the carbon footprint of the rapidly growing digital sector.

The above analysis was intended to show that a long-term and stable energy policy, based largely on low-carbon sources (mainly nuclear energy), combined with proactive investments in grid modernisation (as exemplified by the success of the Linky programme), creates a truly solid foundation for the development of energy-intensive technologies such as artificial intelligence. It is not just a question of the amount of power available, but above all its quality, understood as stability of supply, low carbon intensity and price predictability. Poland, with its historical dependence on coal and current challenges related to grid modernisation, is at a completely different starting point, which makes our path to AI power status much more complex.

The French approach, i.e. strong state involvement in strategic sectors of the economy, such as energy and artificial intelligence, while actively attracting private investment, may be difficult to replicate directly in our circumstances. We have completely different historical conditions, a different energy market structure and we differ in terms of available resources. However, there are important lessons to be learned from the French model regarding the fundamental importance of a long-term strategic vision and the need to create a stable and predictable regulatory environment for investors. France proves that aspirations to be an ‘AI power’ require the status of an ‘energy power’ – or at least a very intelligent and forward-looking energy strategy. Furthermore, France actively promotes the concept of ‘AI sovereignty’, which is inextricably linked to ensuring energy security and independence in this key area.

Key modernisations for Poland

For Poland to realistically aspire to become a European AI leader, profound modernisation of the energy sector is necessary, including production, transmission, distribution and smart system management.

Zero-emission energy sources

It is essential to accelerate the transition of the energy mix towards zero-emission and low-emission sources capable of meeting the growing demand of the AI sector.

• Accelerating investment in renewable energy sources: Exploiting the potential of wind (especially in the Baltic Sea – potential of 18 GW) and solar (over 22 GW installed, target of 29 GW by 2030 according to the National Energy and Climate Plan) requires the removal of administrative barriers, streamlining of permits and a stable investment framework.
• Implementation of the nuclear energy programme: The construction of the first nuclear power plant by 2033 and subsequent units is crucial for stable, zero-emission base energy, which is essential for the AI sector.
• The role of gas and the development of hydrogen technologies: Gas can stabilise the system during the transition period, but it is not a fuel of choice, only a transition fuel. The national hydrogen strategy should be actively implemented, focusing on green hydrogen, which can power gas turbines, industry and transport, and store energy, supporting grid stability.

Modernisation and expansion of the transmission and distribution network

Ambitious plans to develop generation capacity will not succeed without a thorough modernisation of the network, capable of efficiently transmitting and distributing energy.

• Investments in new lines and stations: The PSE Transmission Network Development Plan (over PLN 64 billion for 2025-2034, construction of 4,700 km of new 400 kV lines) must be consistently implemented and potentially accelerated and flexibly adapted to the needs of RES and large consumers, such as AI data centres.
• Strengthening connection capacity: Network bottlenecks should be reduced by streamlining connection procedures and promoting solutions such as cable pooling. PSE plans to make 1,200 MW available to data centres by 2034, but the real needs of AI may be greater.

Smart grids, energy storage and system flexibility

A modern power system that supports the development of AI must be smart, flexible and equipped with energy storage capabilities.

• Implementation of smart grid technology: It is necessary to accelerate the installation of smart meters (currently around 40% in Poland, with leaders already close to 100%) and the construction of advanced network management systems, which are key to the effective management of RES and optimisation of network operation.
• Development of energy storage systems: Investments in battery energy storage systems (BESS), pumped storage power plants and hydrogen technologies are essential for balancing a system with a high share of RES. Energy storage increases the stability and reliability of the system.
• Demand response mechanisms (DSR): Incentives for energy consumers, including AI data centres, to manage their power consumption flexibly should be developed, which will contribute to system balancing and RES integration.

Modernising the Polish energy sector for AI development is a huge challenge that goes beyond building new capacity. It is a profound transformation towards greater flexibility, intelligence and lower emissions. The scale of investment in the grid (over PLN 64 billion by 2034) illustrates this. Success depends on coordinating these investments with the development of generation sources and technologies that support system stability. The example of France shows that Poland needs to catch up to meet the needs of the future and become an AI leader. The success of this transformation depends on attracting private investment through a stable regulatory framework. The current strategic uncertainty is a major obstacle to this. The development of long-term PPAs can stimulate investment, but requires a stable legal and regulatory environment.

Subscribe so you don’t miss any new articles!

Poland vs. France – a comparative analysis

The key difference between Poland and France is their fundamentally different starting points and the degree of consistency in their energy strategies. France is building its position on a solid, low-carbon energy base dominated by nuclear power and has a much more advanced and modernised electricity grid. This situation greatly facilitates the integration of the growing needs of the AI sector, which requires stable and clean energy supplies. Poland, on the other hand, faces a much more complex challenge: it must simultaneously transform its still highly carbon-intensive energy system, modernise and expand its network infrastructure, which is already struggling with stability issues, and prepare for a surge in energy demand from the rapidly growing artificial intelligence sector. This task is comparable to simultaneously renovating an old building, constructing a modern extension and moving in a new, large and demanding family, which is the AI sector.

France is skilfully using its advantage in the energy sector as a geopolitical and economic tool, actively attracting investment in artificial intelligence and positioning itself as a European ‘AI powerhouse’. Its offer is based on ensuring access to clean, stable and relatively cheap energy. If Poland does not accelerate the pace of its energy transition and grid modernisation, it risks not only losing its competitiveness in the race for AI investments, but also deepening its dependence on imports of advanced technologies and potentially energy, or having to bear significantly higher costs for adapting its energy system in the future. Delays in the transition may mean higher energy costs for the Polish AI sector, resulting, for example, from CO2 emission charges, the need to import energy or build costly ‘last-minute’ stabilisation solutions.

Conclusion

Poland’s ambition to become a leader in AI in Europe is a real opportunity, but at the same time it requires a fundamental overhaul of the energy sector. Without decisive and coordinated action in the areas of energy transition, grid development and energy storage, Poland faces an equally real risk of losing its competitive edge in the race for investment in artificial intelligence. The French model shows that a stable and low-carbon energy base is a key asset in attracting and developing the energy-intensive AI sector.

By leveraging its potential in renewable energy and planning the development of nuclear energy, Poland has a chance to build a solid energy foundation for the future. However, the rapid and effective implementation of ambitious modernisation plans, the elimination of regulatory barriers and the creation of a stable investment environment will be crucial. Only in this way can Poland fully exploit its potential, attract global AI leaders and strengthen its position as a leading player on the European artificial intelligence map. Investing in ‘green energy’ for ‘green AI’ is not only a requirement for sustainable development, but a strategic necessity.