AI’s power needs are shifting infrastructure priorities. The next great divide may be between those who can power AI and those who can’t.
getty
In July 2025, much of Europe sweltered under one of its harshest heatwaves in recent memory. In Spain and France, thermometers pushed past 40°C day after day, and the grid groaned under the weight of surging demand. According to Ember, daily electricity use spiked by as much as 14% during the worst of the heat. Just months earlier, a sweeping blackout left millions in Spain and Portugal without power. In that context, the AI boom — with its energy-hungry data centers and ever-expanding compute needs — doesn’t just look like a marvel of progress. It looks like a reckoning.
Every headline about AI breakthroughs floats above an unseen surge — an undertow of electricity and water consumption that’s pulling quietly at the edges of already strained systems. According to the International Energy Agency (IEA), global electricity consumption from data centers is projected to more than double by 2030, with AI being the single biggest driver. But not all countries can keep up. And not all power is created equal.
The Energy Shift Behind AI’s Growth
The AI boom is reshaping more than the tech industry; it’s changing the way the world’s infrastructure looks. Large-scale models, the foundation of today’s generative AI deployments, place enormous demands on electricity and require vast amounts of water for cooling. For context, researchers at MIT say “the computational power required to train generative AI models can demand a staggering amount of electricity, which leads to increased carbon dioxide emissions,” adding that “a great deal of water is needed to cool the hardware, which can strain municipal water supplies and disrupt local ecosystems.”
In this context, clean and affordable energy is no longer a footnote. It’s a fundamental driver of AI readiness.
“Electricity will define the AI landscape just as oil once defined geopolitics,” said Kenso Trabing, founder and CEO of Morphware AI. “The nations that can offer clean, abundant and affordable power will naturally become magnets for AI infrastructure.”
Morphware is part of a new generation of infrastructure companies designing around sustainability from day one. Instead of retrofitting clean energy into carbon-heavy grids, the company built its core operations in Paraguay — home to the Itaipu Dam, one of the world’s largest hydroelectric power sources. That location, according to Trabing, wasn’t just a financial decision but also reflected the company’s belief that energy strategy must be foundational, not an afterthought.
Kenso Trabing, founder of Morphware AI
Morphware AI
“It gave us access to low-cost, renewable energy — which, for us, was critical not just financially but strategically,” Trabing told me. “Over the next decade, I expect we’ll see a realignment where regions with surplus green energy — South America, the Middle East, parts of Africa and certain European hubs — begin to outcompete traditional tech centers like Silicon Valley, simply because the economics of compute won’t add up without sustainable energy.”
Beyond Emissions: Clean Energy As Strategic Moat
While sustainability is a public concern, Morphware’s approach highlights that renewable energy is also a private advantage. Running on hydropower gives the company a big cost advantage — especially compared to similar facilities in the U.S. or Europe. It also protects them from the kinds of fossil fuel price spikes that can send data center bills soaring.
“For us, renewable energy is about more than cutting emissions,” Trabing said. “It brings price stability and flexibility — two things that matter when you’re operating at scale.”
There’s also growing appeal to enterprises trying to reduce their Scope 3 emissions. “As industries adopt AI, they increasingly want to align with providers who can prove sustainability, not just claim it,” Trabing added. “For Morphware, renewables are both a business advantage and a moral imperative.”
Still, building in underdeveloped geographies isn’t without its challenges. Access to skilled labor, latency issues and policy uncertainty remain real constraints — but for some companies, the tradeoff is worth it.
These are not abstract benefits. A location like Paraguay provides natural cooling from hydropower proximity, which helps mitigate water-intensive thermal management. And the ability to scale AI infrastructure affordably in regions that haven’t historically been seen as tech powerhouses opens up a new vision of geographic decentralization.
Location As Leverage
Morphware’s infrastructure now spans both Paraguay and Abu Dhabi — two locations not often paired in tech strategy decks, but increasingly relevant in a climate-constrained world.
“Our decisions are always guided by two principles: Abundant clean energy and global connectivity,” Trabing said. “Paraguay gives us unmatched access to renewable hydroelectric power at Itaipu Dam. Abu Dhabi, on the other hand, provides a strategic gateway between Europe, Asia and Africa.”
That reflects a broader shift: In a world where energy is becoming the primary constraint on AI, compute will migrate to wherever power is cheap, clean, and politically stable. “Together, these locations reflect a strategy to build in energy-rich regions first, then connect those foundations to the broader AI ecosystem.”
But getting there hasn’t been easy, Trabing noted when I asked him about industry challenges the company has faced. “We had to build infrastructure from scratch — roads, transformers, internet agreements — while also bridging cultural and educational gaps,” Trabing said of Morphware’s early days in Paraguay. “The lesson for other builders is that emerging markets demand patience and humility, but the payoff is tremendous.”
Redrawing The Global Compute Map
If clean energy is becoming the defining variable in AI infrastructure, then the map of global technology is about to shift. “I foresee a decentralization of AI infrastructure,” Trabing says. “Instead of everything clustering in the U.S. or China, we’ll see compute nodes spread across regions with energy surpluses.”
This vision has geopolitical consequences. “Politically, energy will become part of AI strategy, with governments treating clean energy not just as a climate issue, but as a competitive necessity,” Trabing added. “Economically, the advantage will shift toward nations that can export ‘compute’ powered by clean energy, much like they once exported oil or manufacturing capacity.”
That framing reframes AI not as a purely technological arms race, but as an infrastructural and ecological one. The real advantage may lie not in who builds the fastest models — but in who can sustain them without destabilizing the grid, the planet, or the communities around them.
As global demand for compute grows, the divide between AI haves and have-nots may increasingly fall along the lines of energy access. Those with abundant, affordable power will build. Those without may struggle to scale, regardless of talent or ambition.
Morphware isn’t alone in rethinking where AI infrastructure belongs. Companies in Iceland, Kenya and beyond are also betting on clean power as the backbone of compute. The real shift underway? It’s not just about who can build, but about who can power it, sustainably and at scale. That’s where the future of AI may lie.
