Future-Proofing AI: How Clean Energy Can Keep Up With Machine Intelligence

Hi everyone,

In the desert heat of northern Nevada, a small plant is quietly tapping the Earth’s deep heat, bringing clean energy back to the surface. It doesn’t look like the future of AI, but it might be. As artificial intelligence reshapes industries, economies, and daily life, its energy appetite is growing faster than anyone expected. And unless the power behind our digital minds evolves just as rapidly, the climate cost could be steep.

How’s where we currently stand, and what’s on the horizon.

The Coming Surge: AI’s Growing Power Appetite

AI is no longer confined to research labs. It’s becoming the backbone of everything from customer service to drug discovery, and with that shift comes a surge in energy demand. The International Energy Agency projects that global data centre electricity use will more than double by 2030, approaching levels roughly equivalent to Japan’s current power consumption. In the U.S., data centres could soon draw more electricity than the steel, cement, and chemical sectors combined!

Much of this energy is devoted to inference (running AI models at scale), but training remains a significant draw. One estimate predicts that training-related energy use will reach up to 134 TWh by 2027, accounting for approximately 0.5% of the current global electricity demand. Combine that with hardware growth (computing capacity is doubling roughly every 3-4 months) and the message is clear: AI isn’t just getting smarter, it’s getting more power-hungry.

The Fossil Trap: Why “Business-as-Usual” Won’t Work

If the grid stays tied to fossil fuels, AI’s growth will turn into a climate liability. In states like Virginia and Texas, utilities are already charging data centres to cover new fossil infrastructure. Meanwhile, some policy proposals still favour conventional power, fast-tracking AI data centres at the cost of weakened environmental oversight.

The danger is a carbon lock-in cycle where rising demand justifies more fossil development, which then hardwires high-emissions energy into the grid. Without planning, the very technology that has the potential to accelerate climate solutions might deepen the problem instead.

Renewable Reinforcements: Solar, Wind & Grid Innovations

The good news is there’s a better path forward. Large-scale solar and wind are expanding fast, and major tech companies including Google, Meta, Microsoft are securing long-term clean energy through Power Purchase Agreements (PPAs) that lock in energy supplies at fixed amounts and costs over a set time frame. Hybrid projects like Nevada’s Stillwater plant, which combines solar, geothermal, and thermal power is a great example of how resource diversity can stabilise supply.

Co-location strategies, like placing data centers near renewable assets are also gaining traction, alongside innovations like battery storage, microgrids, and AI-enhanced energy forecasting. The challenge isn’t just scaling renewables; it’s integrating them into a more adaptive and resilient grid that’s constantly growing.

The Deep Potential: Geothermal’s Underused Power

Of all renewables, geothermal offers something rare: consistent, weather-independent power. Enhanced Geothermal Systems (EGS) are a breakthrough in this field. Using advanced drilling to reach deep, hot rock formations previously out of reach. U.S. estimates suggest EGS could eventually power more than 65 million homes.

Yet geothermal remains currently underused, supplying less than 0.5% of global electricity. Projects like Fervo Energy’s Corsac Station in Nevada aim to change that, scaling from 3.5 MW to 400 MW in the coming years. The potential is huge, especially in heat-rich areas, but deployment is still slowed by skilled labour shortages, policy gaps, and funding issues.

Smart Siting: Designing AI Infrastructure for the Energy Transition

Location matters in all this. Building AI infrastructure near renewable sources and strong grid connections isn’t just greener, it’s smarter. It reduces the need for costly transmission, improves efficiency, and lowers cooling loads. In Australia, Renewable Energy Zones are being developed to match massive clean power generation with industrial and digital demand.

But there are still risks remain. A large Wyoming might benefit from a cool climate and ample electricity, but without stronger renewable commitments, it could lean heavily on natural gas. Today’s siting decisions will shape whether AI becomes a force for decarbonisation, or reinforces the fossil status quo as we keep building.

The Policy Equation: Incentives, Obstacles & What Needs to Change

Clean energy can’t scale fast enough without smarter policy. In the U.S., permitting delays and fragmented investment still slow renewable growth, while geothermal lags due to massive underfunding. Australia’s Capacity Investment Scheme offers a model that uses financial incentives to accelerate clean energy and meet ambitious emissions targets, something we need to see far more of around the planet.

Policy reforms need to include expedited permitting for Geothermal and renewables, unified planning standards, and incentives for 24/7 clean energy procurement that’s aligned with AI’s constantly growing power needs. Workforce development in areas like grid modernisation and geothermal engineering also needs to be treated like a priority.

A Smarter Grid for Smarter Machines

AI represents a massive leap in machine intelligence. But to realise its full promise, it needs to be powered wisely. That means scaling not just the algorithms, but the infrastructure behind them, like grids, policies, locations, and energy systems.

The technologies are ready: wind, solar, geothermal, and smart grids are all advancing. What’s needed now is coordination. Aligning digital growth with sustainable energy. If we get that right, AI won’t just compute faster. It will help drive a cleaner, smarter energy future.

Catch you next week,
James