Are AI and Green Fuels Competing for the Same Resources?

The world, as I see it, is grappling with an unseen energy conflict, a deepening struggle for clean, dispatchable electrons. On one side, I observe the seemingly limitless hunger of AI data centers; on the other, I recognize the crucial rise of green hydrogen and ammonia. This isn't merely about generating more power; I believe it’s a zero-sum game for the very electrons that will define our sustainable future, and AI’s relentless demand is, in my opinion, creating an unexpected premium for solutions like green hydrogen.

AI's Insatiable Appetite: A Grid Under Siege

In my research, I found that global data center electricity demand, largely fueled by AI, surged by a staggering 50% in 2025 alone, reaching approximately 460-490 TWh. This means data centers globally consumed roughly 1.5% of total worldwide electricity in 2024. Projections indicate this will nearly double to 945 TWh by 2030, consuming roughly 3% of global electricity. Some estimates even suggest consumption could reach between 650-1,050 TWh by the end of 2026. The United States, as the world's largest data center market, is particularly impacted; its data centers consumed 176 TWh in 2023, representing about 4.4% of total U.S. electricity. I discovered projections that this could soar to between 325-580 TWh by 2028, accounting for 6.7-12% of total U.S. electricity.

This explosive growth is driving unprecedented capital expenditure. I noted that the combined capital expenditure of five major technology companies—Alphabet, Amazon, Microsoft, Meta, and NVIDIA—surged to more than $400 billion in 2025 and is expected to rise by a further 75% in 2026. Specifically, hyperscalers like Alphabet, Amazon, Microsoft, and Meta planned to invest over $350 billion in data centers in 2025 and approximately $400 billion in 2026. By 2026, I learned that the total data center CapEx across these four firms alone is projected to hit a record $680 billion.

Individual companies are making massive commitments. Google, for instance, plans to spend between $175 billion and $185 billion in capital expenditures in 2026, nearly doubling its 2025 spending. I found that Google is investing $40 billion through 2027 to build three new data centers in Texas and also announced new facilities in Minnesota, Germany, and a $15 billion AI hub in Visakhapatnam, India. Microsoft, meanwhile, has committed a sweeping $80 billion investment into building and expanding AI-optimized data centers through 2028, with over $80 billion allocated for fiscal year 2025 alone. My research shows Microsoft is planning new data centers across North America, Europe, Asia, and Africa. Amazon Web Services (AWS) spent $96.5 billion on data center capital expenditure in 2025 and has budgeted an astonishing $200 billion for 2026, making it what some call the most aggressive private infrastructure buildout in history. AWS is investing up to $50 billion to expand AI and supercomputing infrastructure for U.S. government customers by 2026, and is also developing major new campuses in Indiana, Mississippi, North Carolina, Pennsylvania, and Virginia. This demand isn't just large; I see it as continuous, high-density, and having near-zero tolerance for interruption, making reliability a paramount concern. I also noted that occupancy rates in data centers are projected to climb from 85% in 2023 to over 95% by late 2026, underscoring the urgency.

The Grid's Breaking Point and the Scramble for Power

The existing grid infrastructure, designed for predictable power, is visibly struggling under this new burden. Regions like Northern Virginia, a major AI hub, are experiencing acute grid stress, with AI-driven energy demand outpacing available capacity. I learned that in July 2024, a voltage fluctuation there even triggered the simultaneous disconnection of 60 data centers, leading to a 1,500-megawatt power surplus that required emergency adjustments to prevent cascading outages. This bottleneck is forcing data center developers to delay projects or resort to building their own onsite natural gas generation, with orders for gas turbines soaring by 70% in 2025.

Beyond Northern Virginia, I’ve observed similar strains globally. In Ireland, data centers now consume 22% of the country's total electricity, a figure projected to reach 32% by 2026. Both the UK and the US are seeing data centers consume 6% of their national electricity supply, and in the UK, the queue to connect to the national grid grew by a staggering 460% in the first half of 2025. The Electric Reliability Council of Texas (ERCOT) projects that peak summer power demand could approach 145 GW by 2031, up from 85 GW in 2024, with over half of this new demand (about 32 GW) projected to come from data centers, including cryptocurrency miners.

While tech giants have aggressively procured renewable energy through Power Purchase Agreements (PPAs)—accounting for 43% of all global clean energy PPAs in 2024, driving prices up by 35%—these agreements often don't guarantee the 24/7 firm power AI needs. Grid congestion, extreme weather, and slow interconnection timelines mean that even with clean energy contracts, continuous availability is not assured. I found that the tech sector accounted for around 40% of all corporate PPAs signed in 2025, and is now also a major source of momentum for the nuclear and advanced geothermal industries. The pipeline of conditional off-take agreements between data center operators and small modular reactor (SMR) nuclear projects, for example, has grown from 25 gigawatts at the end of 2024 to 45 gigawatts today. Companies like Talen Energy are already providing nuclear power for AWS, and firms like Bloom Energy (fuel cells) and GE Vernova (turbines, grid infrastructure) are seeing significant growth.

AI's Unexpected Alliance with Green Hydrogen

This reliability crisis is precisely where green hydrogen, and its derivative green ammonia, enter the fray as an unlikely ally. Produced via electrolysis powered by renewables, green hydrogen offers long-duration, dispatchable, and clean backup power that batteries simply cannot provide for extended outages. Historically, green hydrogen has struggled with economic viability, but AI's urgent need for uninterrupted, clean power is changing the equation. AI data centers are becoming a critical market, providing the stable, high-value demand necessary to de-risk and scale green hydrogen projects.

I've been tracking the impressive growth in green hydrogen initiatives. Over 1,500 green hydrogen projects have been announced across more than 70 countries, representing over half a terawatt of planned electrolyzer capacity valued at around $680 billion. While only $75 billion of this has reached a final investment decision, the momentum is undeniable. The NEOM Green Hydrogen Project in Saudi Arabia, for instance, is the world's largest under construction, backed by Air Products, ACWA Power, and NEOM. It will utilize 4 GW of wind and solar power to run 2.2 GW of electrolyzers, producing approximately 600 tonnes of hydrogen daily (equivalent to 1.2 million tonnes of green ammonia per year) from an $8.4 billion investment, and is on track for a 2026 launch.

In the U.S., I found 76 green hydrogen projects planned, backed by $36 billion in investment, with states like Texas, Louisiana, Alabama, and California leading the charge. Plug Power's St. Gabriel plant in Louisiana began commercial operations in April 2025, producing 15 tons of green hydrogen daily, and the company aims to scale North American production to 500 tons per day by the end of 2025. Their Genesee County project in New York, a 120 MW facility, is expected to produce 74 tonnes of green hydrogen per day when operations begin in 2026. The Intermountain Power Project in Utah, in partnership with Mitsubishi Power, aims to operate on a 30% hydrogen blend by 2025, progressing to 100% hydrogen by 2045. AES Corporation and Air Products are also developing one of the largest green hydrogen facilities planned in North America, located in Texas. Internationally, the North Sea Offshore Hydrogen Hub, powered by offshore wind, is among the most anticipated projects of 2026, aiming to supply green hydrogen to Germany, Denmark, and the Netherlands.

Beyond the Grid: The Water-Energy Nexus and Integrated Solutions

My analysis reveals that the competition extends beyond just electrons; it encompasses water, creating a critical water-energy nexus. AI data centers, particularly those in hotter climates, require immense amounts of water for cooling. I’ve learned that this water-dependent cooling already strains local ecosystems across Northern Virginia, Ireland, Singapore, and parts of Western Europe. While green hydrogen production is often criticized for its water demands, I found that the chemistry dictates 1 kg of hydrogen requires 9 kg of water. After accounting for purification, cooling, and other auxiliary systems, most commercial green hydrogen plants consume 20-30 liters of water per kilogram of hydrogen. For a 100 MW plant with high utilization, this could mean 400,000 to 600,000 liters of water per day. However, when viewed through a life-cycle lens, green hydrogen is often more water-efficient than fossil fuel-based energy generation; for example, coal-fired power stations typically evaporate about 1.9 liters of water per kWh in their cooling towers. Crucially, I've noted that large-scale green hydrogen projects, such as NEOM, are opting for renewable-powered desalination or using treated municipal and industrial effluent, mitigating freshwater impacts. Moreover, I was interested to see that some new data centers, like Google's facility in Wilbarger County, Texas, are incorporating advanced air-cooling to eliminate operational water use, minimizing their local grid and water impact.

This dual resource challenge also presents an opportunity for synergistic solutions. I believe we will increasingly see data centers co-locating with renewable energy projects and hydrogen production facilities, creating integrated energy hubs. Imagine a wind farm powering both a data center and an electrolyzer, with the green hydrogen providing backup power for the data center during lulls in wind production. Waste heat from data centers could potentially be harnessed for district heating or even for specific stages of hydrogen production processes, further enhancing efficiency.

Geopolitics and the Future of Clean Electrons

The intensifying competition for clean, dispatchable electrons also carries significant geopolitical implications. I've been following what experts are calling a "triple transition"—the simultaneous transformation driven by advances in AI, changing energy systems, and an accelerating geopolitical realignment. Strategic control over compute, data, and digital infrastructure has, in my opinion, acquired the same geopolitical significance as shipping lanes or energy pipelines. The U.S.'s sustained AI leadership, for example, may depend on its ability to effectively address these energy demands, potentially outpacing rivals who are also keen to meet them.

I've observed that governments worldwide are recognizing this. The European Union, through its Digital Decade Policy Programme (DDDP), emphasizes renewable energy use and sustainable design for data centers to achieve climate neutrality by 2030. The EU AI Act even mandates environmental impact assessments and energy efficiency standards for high-risk AI systems. An analysis by the Kiel Institute for the World Economy projects that data centers associated with the EU’s "AI Continent Action Plan" could consume up to 168 terawatt-hours of electricity by 2030, roughly equivalent to Poland’s current power demand or about 5% of overall EU consumption. In response, the European Commission will be putting forward a Data Centre Energy Efficiency Package in Q1 2026, aiming for carbon-neutral data centers by 2030. Similarly, countries like China and Gulf nations such as Saudi Arabia, through initiatives like Vision 2030, are investing heavily in AI and associated energy infrastructure, understanding the strategic advantage that reliable, clean power for AI confers.

What This Means For Investors/Entrepreneurs/Professionals

For investors, I see compelling opportunities in companies providing grid infrastructure upgrades, energy storage solutions, and advanced cooling technologies for data centers. Firms specializing in SMR nuclear technology or advanced geothermal are also poised for significant growth, given the tech sector's increasing interest. Furthermore, I believe that investments in green hydrogen production, particularly those with secured off-take agreements from industrial users or data centers, represent a strong play. Companies like Plug Power, Air Products, and those involved in large-scale projects such as NEOM are worth watching.

Entrepreneurs should focus on innovative solutions at the intersection of AI, energy, and water. This could include AI-driven grid optimization software, modular data center designs that integrate renewable energy and hydrogen fuel cells, or advanced water recycling and air-cooling systems for data centers. Developing technologies that can convert data center waste heat into usable energy is another promising area.

Professionals in energy, technology, and policy must develop interdisciplinary expertise. Energy planners need to understand AI's granular power demands, while tech leaders must grasp grid limitations and the economics of green fuels. Policymakers face the urgent task of creating regulatory frameworks that incentivize clean energy deployment for AI without stifling innovation, promoting transparency in energy and water usage, and fostering international collaboration on energy security and sustainable digital infrastructure.

Bottom Line

I believe the battle for clean electrons between AI and green fuels is reshaping our energy landscape, creating both immense challenges and unprecedented opportunities. Securing a sustainable future requires strategic investments in reliable, clean power sources and a holistic approach that addresses the energy, water, and geopolitical dimensions of this unfolding technological revolution.