The AI data center boom is quickly becoming an energy story as much as a technology story. A new report from the International Energy Agency says the world will spend $580 billion on data centers this year, which is $40 billion more than will be spent finding new oil supplies.
That comparison captures the scale of the shift now underway. Generative AI needs large computing facilities, and those facilities need power. The open question is how much of that demand can be met by renewable energy, and how much pressure the buildout will place on electrical grids that are already under strain.
Why the spending number matters
The $580 billion figure is not just a large infrastructure number. It shows how central data centers have become to the global economy. In the discussion around the International Energy Agency report, TechCrunch’s Equity podcast framed the comparison with oil spending as especially relevant because of concerns that generative AI could accelerate climate change.
Data centers are power-intensive by design. As companies build more of them for AI workloads, electricity demand rises with them. That creates a practical problem: even when companies can pay for buildings, chips and land, they still need reliable grid connections and enough power to operate at scale.
The report’s findings also point to where the pressure may be concentrated. According to the discussion, half of the electricity demand will come from the U.S., while the rest will be a mix of China and Europe. Many of the data centers are expected to be in or near cities, including areas with populations of a million people, roughly.
That location pattern matters because grid connections become more complicated near dense population centers. Large projects need connection pathways, and those pathways can be harder to secure when they compete with existing demand from homes, businesses and public infrastructure.
Why solar is getting attention
Renewable energy is not being discussed only as a climate preference. In the podcast conversation, Kirsten Korosec pointed to solar as a potential upside of the data center buildout because it can be a practical answer to regulatory and cost issues.
The reasoning is straightforward. It can be easier to get a permit to install solar panels next to a data center than to rely entirely on more complex grid upgrades or other power sources. If solar can be built close to the computing facility, it may help reduce some of the friction that comes with placing a large new electricity user on an already taxed grid.
Rebecca Bellan also noted that renewables may become a focus because they make business sense, not simply because of environmentally friendly policies. For developers trying to get projects powered, the fastest or most workable energy source can become the most important one.
That creates a potential opening for companies working on:
- renewable energy projects linked to data centers
- data center design that reduces the global emissions component
- technology that helps large computing sites manage electricity demand
- new ways to connect power generation with AI infrastructure
The key point is that the AI boom may pull renewable energy deeper into the center of infrastructure planning. Solar is not just a cleaner option in this context. It may become part of the basic business case for getting data centers approved, connected and running.
Microgrids could become part of the answer
The source discussion also highlighted Redwood Materials’ new business unit, Redwood Energy, as one company to watch. Its approach uses old EV batteries that are not quite ready to be recycled, then turns them into microgrids aimed specifically at AI data centers.
That model is important because it addresses one of the major concerns around the AI data center buildout: whether the grid can handle the additional load, especially during periods of high demand. Kirsten Korosec pointed to the middle of the summer as one example, and cited places like Texas that have rolling brownouts and blackouts.
Microgrids do not remove the broader energy challenge, but they can change how a data center interacts with the grid. If stored power can help support operations during strained periods, it may reduce some of the pressure that would otherwise fall on local electrical systems.
The bigger question is whether more companies will pursue similar models. If Redwood Energy is part of a wider pattern, the data center boom could encourage investment in companies that combine batteries, renewable energy and grid management. If it remains more isolated, the pressure on grid operators could be harder to ease.
The buildout still faces hard questions
The scale of announced data center spending is enormous. OpenAI has said it has committed $1.4 trillion to building data centers. Meta has committed $600 billion. Anthropic recently announced a $50 billion data center plan.
Those commitments show how aggressively major AI companies are preparing for future demand. They also raise a practical question: how much of the planned infrastructure will actually get built? The plans require huge amounts of spending, and the source discussion noted ongoing questions about OpenAI’s revenue compared with the trillions of dollars of capital commitments it has for the next decade.
Financing may also pull governments further into the debate. The discussion referenced controversy over OpenAI’s CFO saying, “The government should backstop our loans to build these data centers.” She later said she did not mean backstop and called it a poor choice of words. The discussion also noted that OpenAI appears to have been asking for an expansion of tax credits from the CHIPS Act.
That means the data center boom may not be decided by technology companies alone. Over the next few years, governments may have to consider how much support, if any, should go toward the infrastructure behind AI.
What the renewable energy question really means
Asking how much of the AI data center boom will be powered by renewable energy is really asking several questions at once. Can solar be built close enough and fast enough to support new projects? Can microgrids reduce stress on electrical grids? Can startups build useful tools around renewables, batteries and data center design?
There is no single answer in the source material. What is clear is that AI infrastructure is moving from an abstract technology issue into the physical world of power lines, permits, batteries, land and grid capacity.
If renewable energy becomes the practical path for many new data centers, the boom could accelerate investment in solar and related technologies. If it does not, the same boom could deepen pressure on already taxed electrical grids. Either way, the energy system is now part of the AI story.