Let’s start with the basics. What is Fossefall? What are you trying to achieve?
Øyvind L. Vesterdal. – “Fossefall is a Nordic and European alternative to the global giants of artificial intelligence. Access to computing power for AI is set to become a scarce resource very quickly. In the future, it will be an exclusive commodity. Today, the bulk of this capacity is concentrated in the United States. The rest is mainly in China, with some in the Middle East. In Europe, we simply do not have enough. In the current geopolitical context, and given all the sovereignty issues at stake, Europe must have its own artificial intelligence capabilities. This is precisely Fossefall’s objective. Today, our main constraint is neither demand nor funding. For the first time in my professional life, I have more clients and greater access to capital than I have capacity to sell to them. Our ambition is therefore to harness Nordic hydropower to serve the European – and, ultimately, global – market for computing capacity for artificial intelligence.
But, in practical terms, what are you building? Data centres dedicated to AI?
“Yes, but we prefer to talk about ‘AI factories’. These days, most players build a data centre, find a client – usually an American hyperscaler, sometimes a major Chinese player – and then sell that computing capacity to them. That client then resells that capacity to its own users. In reality, most data centre operators only own the building. They don’t own the machines that create the value. Yet, if you look at any factory, most of the value comes from the equipment inside it, not the walls. For us, it’s exactly the same logic. We don’t just own the data centre: we also own the machines inside it. This is one of the key factors that sets Fossefall apart.
Norway is often described as Europe’s battery.
I imagine you mainly buy Nvidia chips.
“That is indeed what most of our customers are asking for today. They mainly want to use Nvidia chips, even though there are other players in the market. I am, in fact, eagerly awaiting the emergence of a genuine European alternative. But, to be realistic, any company building AI infrastructure today is essentially buying American chips. Nvidia remains the benchmark. There are also AMD, Groq, SambaNova and several other suppliers gradually entering the market. The challenge isn’t just about having good technology. You also need to be considered ‘bankable’, that is to say, credible enough to convince investors and financial institutions. Today, when you approach the financial markets, you very quickly realise that the American offering is practically the only one that meets these criteria.
The other pillar of your business model is, if I understand correctly, your access to Nordic electricity, which is much cheaper than in the rest of Europe. This is one of your main advantages.
“Yes. Norway is often described as Europe’s battery. This view has long been held in the Nordic countries. Northern Europe’s largest hydroelectric reservoir is about an hour’s drive from my home. Northern Norway and Sweden still have an abundance of electricity. At certain times of the year, prices even turn negative. Today, where I am, the outside temperature is between 11 and 12°C. You obviously don’t find these conditions in Paris or in southern Europe. However, AI plants require even more cooling than traditional data centres. Setting up in a region where electricity prices are predictable, low and where supply is stable is therefore a considerable advantage. In many countries, the price of electricity has become directly linked to that of gas. In the Nordic countries, this dependence is much lower and energy remains much more readily available.
Luxembourg has all the necessary ingredients: financial expertise, the ability to structure complex transactions, and recognised know-how in raising the capital required for this type of project.
Do you, for your part, have access to 100% renewable electricity? Is this one of Fossefall’s selling points?
“Sustainable development is at the heart of our approach. Of course, we all want to contribute to a greener planet. But, beyond this environmental aspect, it is also a perfectly rational choice from an economic point of view. By relying on renewable energy, we are much less exposed to hydrocarbons and to the volatility of fossil fuel prices. In Norway, around 98.5% of electricity is generated from hydropower. The remainder comes mainly from wind power, supplemented by electricity exchanges with neighbouring countries. In Sweden and Finland, the energy mix is based on hydroelectricity, wind power and, to a certain extent, nuclear power.
These days, customers often place greater importance on price than on CO2 emissions or sustainability. At Fossefall, on the other hand, this aspect is an integral part of our identity. We are also one of the few companies developing new energy-generation capacity alongside our AI facilities. In practical terms, we are working on projects that combine renewable energy generation with battery storage directly linked to our infrastructure. At present, this remains an added bonus. But I am convinced that, within a few years, it will quite simply be an essential requirement for operating this type of infrastructure in Europe.
In Luxembourg, we also have an ‘AI Factory’. I imagine, however, that this isn’t exactly the same as your project. Could you explain the difference? Why isn’t your initiative part of a wider European project centred on ‘AI Factories’? Why are you developing this activity as an independent company?
“I’ll start with the second question. Fortunately, I’m not a lone entrepreneur. I’m surrounded by excellent co-founders, strong shareholders and institutional investors, the majority of whom are European and Nordic. They include both entrepreneurs and industrialists. I’m seeing a real shift in awareness across Europe. Many people still think that artificial intelligence is essentially all about software. In reality, the infrastructure that enables this artificial intelligence to function is becoming one of the next major infrastructure asset classes in Europe. This is precisely why Luxembourg plays such an important role. Here, you have all the necessary ingredients: financial expertise, the ability to structure complex transactions and recognised know-how in raising the capital essential for this type of project.
Only the energy and oil and gas sectors currently attract comparable levels of capital.
This naturally brings us to the company you set up in Luxembourg at the end of June. What is its purpose? Why set up a company in Luxembourg when you could also raise funds in the Nordic countries?
“We are, of course, continuing to raise capital directly in the Nordic countries. But we are talking here about an asset class that requires considerable sums. In my view, only the energy or oil and gas sectors currently mobilise comparable volumes of capital. If we look at current investment in data centres and artificial intelligence, it already exceeds that allocated to many energy projects. Although we are very pleased with the funding secured in the Nordic countries, I regard this project as a European venture far more than a purely Nordic one.”
Governments, institutional investors and industry partners are all seeking to address the same challenge, each from their own perspective. Luxembourg is one of the very few places where favourable public policy, a stable and predictable legal framework, and an ecosystem that is fully understood by institutional investors all come together. In my view, this makes it the ideal hub for driving forward European projects.
Could you please state how much you wish to raise and over what period?
“Over the next few years – without going into too much detail about the timetable – we plan to finance around 30 billion dollars’ worth of GPUs. I should point out that this relates solely to the financing of graphics processing units. It does not include the financing of the data centres themselves. We are therefore talking about a project of considerable scale. We have already raised funds in the Nordic countries to finance the holding company, which will enable us to provide the necessary equity for our first project. But behind this lies an extremely large portfolio of projects to be developed. We probably have one of the largest portfolios of electrical capacity – in megawatts – in Europe, even though our company is still very young. For us, this is just the beginning of the journey.
When data centre developers present their projects, they often highlight a proprietary technology. Is this also the case for you? Or does your competitive advantage lie primarily in your access to low-cost energy in the Nordic countries?
“I would highlight two or three key points. From a technological perspective, we obviously do not manufacture our own chips. Cooling solutions already exist on the market, and our philosophy is simply to use the best available technologies. What really sets us apart is our GPU platform. It allows us to be much more than just an infrastructure provider. Simply owning the chips isn’t enough. You also need the software layer that enables you to commercialise this computing power. This is what positions us alongside players such as Amazon and Microsoft, who now sell artificial intelligence services. Whether it’s a small business, a public body or a government, anyone will be able to purchase computing capacity via Fossefall’s GPU platform. That is our first major point of difference. The second is our access to energy, which remains the sector’s main constraint today. Finally, we are fully integrated. We manage the electricity supply, the data centres and the GPUs that power our platform. Very few companies are willing to integrate the entire value chain, as this requires expertise spanning every link in the chain. This is precisely one of our key strengths.
We do not want to dedicate our infrastructure to purely recreational purposes (…) Everything we do must serve a purpose for society.
For several years now, hyperscalers have realised that their customers want to keep their data as close as possible to where it is generated. That is why they have been expanding their network of data centres across Europe. Where will your AI facilities be located?
“Our projects will mainly be based in the Nordic countries. If Europe were a single country and you were free to choose where to build your infrastructure, you would naturally set it up in a region with a cold climate, whilst benefiting from excellent infrastructure and good connectivity. That is exactly what the Nordic countries offer us. That is why we have chosen this region.
Companies with access to abundant, low-cost energy have often chosen to put this capacity to use in the cryptocurrency industry. Why haven’t you followed this path?
“It’s a very clear strategic decision: we will not be getting involved in cryptocurrencies. Nor do we want to dedicate our infrastructure to purely recreational uses – I often joke that we don’t want to power cat videos. Everything we do must serve a purpose for society. Our aim is for the infrastructure we develop to be regarded as critical infrastructure. That is why we primarily target governments and companies whose activities, in our view, fall into this category. Today, when a new data centre is announced for entertainment platforms such as TikTok or other similar services, there is often strong public opposition. Conversely, when citizens understand that this infrastructure meets an essential need for society, it is much more readily accepted. This is precisely the approach we have chosen to take.
Finally, could you give us some figures about your company?
“I would say that we represent a rather rare combination of a start-up and a company that is already highly structured. Both the founders and the management team are highly experienced individuals. We have previously set up and grown companies. We have already put in place the governance systems, processes and quality standards that institutional investors generally expect from a much more mature company. We built Fossefall from the outset as a company capable of developing predictable and bankable infrastructure projects. We are not a start-up software company founded in a garage. We are, first and foremost, an infrastructure company. We are not yet a large company, but our ambition is clearly to become one of the leading players in AI infrastructure in the Nordic countries, and subsequently in Europe. We currently have around thirty employees. The company has been in existence for ten months. We have also just finalised our first Series A funding round, raising around 500 million Norwegian kroner, or nearly €50m. On a global scale, this is still a modest company. However, in the field of artificial intelligence infrastructure, we believe we now have all the necessary ingredients to become a major player on the international stage.”
