The US grid simply cannot interconnect fast enough to meet skyrocketing demand. Are we witnessing a permanent structural shift toward off-grid, behind-the-meter power, or is this a temporary gap that utilities will eventually close?
I think you're absolutely right that there's an inherent capacity challenge in the US grid. The demand is coming from multiple directions at once: aging power plants across all types needing replacement, the general electrification of many sectors, and now on top of that, the AI data center supercycle, which is simply unprecedented. I've been in the power industry for 30 years and I've never seen anything like it. Colleagues who lived through the Enron era say even that wasn't like this.
But to your question about permanence: I don't think the behind-the-meter plants being built today will become stranded assets in 15 years. Even if grid connections eventually reach data centers, these modular plants can shift to serving what we call balancing power, supporting a grid that will by then carry a much larger share of renewables. They are truly flexible assets, a Swiss Army knife in many respects. The current moment is about urgency; the long-term case is about adaptability.
Beyond data centers, reshoring in the US is accelerating, chip fabs, EV plants, LNG terminals. Is that industrial wave showing up meaningfully in your pipeline yet?
Honestly, right now all players in our space are somewhat overwhelmed by the data center boom, and I think everything else is getting a little less headline attention. One reason data centers dominate is speed: traditional utilities and industrial customers have buying cycles where the journey from first inquiry to contract can take years. With data centers, we're talking months. That said, other customer segments, including utilities, are learning that if they want equipment they need to start moving faster, and many are adapting quite well to the new reality.
What I'd highlight is that the three things really driving data center site selection are fiber connectivity, access to gas, and available grid capacity. Northern Virginia is effectively closed to new data centers because the grid simply cannot absorb more load. That dynamic is pushing developers toward new geographies, which in turn opens opportunities for us across a broader footprint of sites.
Gas demand is accelerating sharply. How do you make the case for gas engines to customers and policymakers who still see gas as part of the problem?
I'm a power plant engineer who has done extensive modeling work, so I'll be direct: onshore wind and solar have the lowest levelized cost of energy and there is no serious question about that near population centers. The real question is which technologies provide the most carbon-efficient firm power when renewables are not generating. Large-scale nuclear is probably too slow for the current cycle, and SMRs are not yet commercially proven at scale. That leaves gas as the practical answer for now.
What makes our position more durable is fuel flexibility. We have already run tests with natural gas and hydrogen blends with WEC in Michigan, and our engines can also run on ammonia and ethanol. So whatever cleaner fuel becomes widely available and affordable, we are positioned to transition with it. The goal is to ensure the assets we sell today are never stranded, and that future-proofing is a genuine technical capability, not just a sales point.
When does hydrogen actually become an operational reality for US customers rather than just a selling point?
Within the current political environment, some of the funding and pilot programs that were driving early hydrogen development have been put on hold, so the timeline has become less certain. My honest view is that 100% hydrogen combustion in thermal generation is still some way off. Blending will come first. The full green hydrogen cycle, converting solar or wind to hydrogen and then back to electricity, is not particularly efficient, and there are other parts of the industrial value chain that will likely benefit from hydrogen earlier than power generation. We are ready for blending now, but a large commercial market for 100% hydrogen thermal generation will take time to materialize.
Competitors like GE, Caterpillar, and INNIO are chasing the same orders. As this market becomes more crowded, what is your durable competitive advantage?
I think we're seeing a genuine structural shift in what customers actually need, and our technology fits that shift better than large monolithic power plants do. A single 500-megawatt combined cycle unit does not work well for a data center being built in stages. What works is modularity: engine blocks of roughly 100 megawatts that can be added incrementally as the campus grows. Our plants are also highly flexible in operation, which matters increasingly on the utility side as more renewables enter the grid and balancing power becomes the premium product, not baseload.
One point I want to highlight that often surprises people: our cooling system is fully closed-loop, like a radiator. We joke internally that our power plant uses about as much water as however often the staff visits the restroom. In regions with serious water constraints, which is a growing concern across the US, that is not a minor detail. It is a genuine operational advantage that our larger turbine-based competitors simply cannot match.
You are moving beyond selling equipment into operating plants under long-term agreements. How does that shift your relationship with customers and your own risk exposure?
It is a positive challenge and one we prepared for deliberately. We already have more than six gigawatts planned in the US, and several years ago, as we saw balancing power demand rising, we began investing in local infrastructure ahead of the curve. We opened a remote monitoring and expertise center in Houston, which can connect to all US plants while keeping data onshore, and last summer we opened a Houston Logistics Center as a satellite warehouse to serve customers faster and with greater supply chain security. Those investments are paying off now.
The other dimension is workforce. As these plants come online and run on baseload for data center customers, they require skilled maintenance and operations staff. We are actively hiring US-based engineers, and collaborating with universities and vocational schools to build that pipeline. It is not glamorous work, but it is essential, and being ahead of it is part of what makes a 10-year operating agreement something we can actually deliver on rather than just sign.
Washington is pushing hard on energy dominance, more gas, faster permitting. Does that make your job easier or more complicated?
I think the honest answer is that local US supply will never fully meet US demand in this cycle, and that applies to us as it does to Siemens and Mitsubishi. Our main factory remains in Finland, and we are transparent about that. What we have done is invest seriously in US presence: the logistics center, the monitoring expertise center, and a growing field services team on the ground. The policy push for faster permitting and more gas infrastructure is genuinely helpful for the overall market we serve. We want to be as present as we can be and to support US energy objectives with our current and future business, and we think we are well positioned to do that.
Looking five years out, what are the two or three things you most want to have achieved?
First, ensuring that the solutions we are selling today are never stranded assets. That means continuing to invest in fuel flexibility and making sure our plants can evolve as the energy mix changes. Flexibility is not a marketing concept for us; it is an engineering commitment. Second, energy security is increasingly urgent, and recent events, including the blackout in Spain, are a reminder that even very developed grids can fail. Our modular plants can support grid restoration and resilience in ways that large centralized generation cannot. We want to be an integral part of how countries plan for that. The company has been around for more than 190 years, and we intend to be a meaningful part of the energy system for another 190.
Beyond the immediate pressure from data center demand, how do you see Wärtsilä’s engine technology fitting into the bigger picture - supporting industrial growth and long-term energy security?
I think the data center story is really part of a much bigger electricity story. Demand is growing from a lot of directions at once - data centers, yes, but also electrification, manufacturing, broader industrial growth, and the need to replace ageing generation.The challenge is that the grid simply cannot expand at the same speed as all of that demand. That is where flexible generation becomes so important. For a data center, this technology can solve an immediate power need. But it is not limited to one purpose. Over time, the same flexible generation capacity can support broader industrial demand or serve the grid as balancing power, especially as more renewables come online. Our engines are modular - they can be built in phases, scaled up as demand grows, and adapted as the energy mix changes. And critically, they are not locked into any single fuel. They can run on natural gas today and transition to hydrogen blends or other sustainable fuels as those become available.
For me, that is the key point.
The power system is going to need infrastructure that is reliable today but also adaptable in the future. That makes this kind of infrastructure a practical answer to today’s urgency and a genuine long-term part of energy security.
As electricity becomes increasingly central to modern life and economic growth, what role can reliable, flexible power play in supporting grid stability, energy resilience, and sustainable societal development?
Reliable electricity is becoming one of the foundations of modern society. Healthcare, financial systems, communications, logistics, public services, national security and industrial activity all depend on electricity being available when it is needed. Data centers are often talked about in the context of AI, but their role is much broader than that. They support the digital systems that many essential services and industries now rely on.
So the question is not only how we meet one source of demand. It is how we build a power system that can support the society and economy we are becoming.
