What fascinates me most is not the algorithms, but the fact that the future of technology is becoming brutally tangible. Cables, cooling loops, substations, storage. And this is where Scandinavia finds itself at a pivotal moment: uniquely positioned, structurally underbuilt, and increasingly impossible to ignore.
Everyone talks about AI. Few talk about what breaks first.
Artificial Intelligence is often described as software. Code. Models. Intelligence in the cloud. That framing is misleading. AI is not floating in the ether — it lives inside massive physical halls filled with servers, power infrastructure, cooling systems, and redundancy layers. A single hyperscale facility can span hundreds of thousands of square meters.
Switchis a clear example of that: their data center campuses are not IT products, they are industrial-scale assets — closer to power plants than software companies, and it looks really massive.
By 2025, of all the capacity available 33% of global data-center is expected to be dedicated to AI workloads, while total electricity demand from data centers is projected to grow by 16% year over year. That growth does not stress code first — it stresses power, cooling, space, and resilience.
This is where most investors misread the opportunity. The most durable value is not created by the AI models themselves, but by the solutions to the problems AI creates: power density, cooling limits, latency, load volatility, grid interaction, and sustainability.
These are not optional upgrades. They are existential requirements. AI doesn't scale because software improve, it scales only when the halls behind it can.
2) Where the world is building — and why the Nordics remain underexposed
Global hyperscale capacity is concentrated in a few dominant hubs. Northern Virginia Data Center Alley is the world’s most dense data-center cluster. Frankfurt serves as Europe’s digital backbone. Singapore continues to scale despite severe climate and energy constraints.
And yet, the Nordic region — with cold climate, political stability, robust grids, and one of the cleanest energy mixes globally — lacks comparable hyperscale clusters. From an investment standpoint, that imbalance is striking. Structural demand is rising faster than supply, and geography still matters when energy becomes the bottleneck.
3) The problems AI creates — and why they are investable
Modern AI data centers operate under conditions that traditional infrastructure was never designed for. Each challenge represents a focused investment thesis:
- Extreme power density AI racks demand 30–80 kW per rack, compared to 8–15 kW in legacy environments. This drives demand for advanced power electronics, grid-interface solutions, and integrated energy-storage systems.
- Cooling beyond air Air cooling has reached its physical limits. Liquid cooling — direct-to-chip and immersion — can be up to 3,000× more effective, enabling higher compute density and lower failure risk. Companies mastering this layer sit at the heart of future data-center design.
- Load volatility and resilience AI training produces sharp power spikes. This creates demand for modular UPS systems, fast-response power management, and intelligent load balancing — all mission-critical and contract-driven.
- Non-negotiable sustainability Water efficiency, carbon intensity, and energy transparency are no longer branding exercises. They are contractual requirements from hyperscalers, governments, and institutional capital.
4) Why this beats chasing the “Next AI Unicorn”
There is a clear distinction between hype capital and infrastructure capital. AI software companies may offer exponential upside — and exponential volatility. Infrastructure providers solving AI’s physical constraints play a different game:
- Long-term contracts (often 10–20 years)
- High-quality counterparties (hyperscalers, utilities, public entities)
- Predictable, often inflation-linked cash flows
- Structural demand independent of economic cycles
This is growth with discipline. Exactly what long-term capital is searching for in an increasingly unstable macro environment.
5) My conviction: invest where the future is locked in
What fascinates me most about this shift is how tangible it is. AI and blockchain may be digital by nature, but their execution is brutally physical — cables, cooling loops, substations, storage systems. Real assets, real constraints, real engineering.
What I enjoy even more is the ability to take ideas born in conversations, whiteboard sessions, and late-night pattern spotting — and move them into execution through
Burhall & Partners. From there, we can structure, validate, and scale projects together with
the finest operators in industry, while bringing in the capital required to actually build what the market needs.
My thesis remains simple: don’t only invest in what captures attention — invest in what the future cannot function without. The Nordic region has a unique opportunity to become a cornerstone for next-generation data-center infrastructure by backing companies that solve power, cooling, and operational resilience at scale.
The returns are attractive. The contracts are long. And the demand is not cyclical — it’s structural.
Whether this article marks the beginning of a new project remains to be seen. But history suggests that’s often how things start — especially considering we are already deeply involved in battery energy storage, where power, resilience, and infrastructure thinking are not theory, but daily practice.
■
