Venture capitalists are making increasingly large investments, investing more than $500 billion in AI startups over the past five years.
But according to a report from Sightline Climate, the smartest AI investments these days may be in energy. Researchers found that up to 50% of announced data center projects can be delayed. One of the biggest causes is access to power.
Of the 190 gigawatts worth of data centers the company is tracking, only 5 are under construction. About 6 gigawatts of data center projects in Sightline’s database came online last year. An even larger proportion (about 36%) see their schedules falling behind schedule in 2025. The delay could eventually trickle down and impact large corporations and other companies that use AI in their business.
This supply-demand pressure is an opportunity for investors. Here’s why:
Tech giants like Google and Meta have invested significant portions of their balance sheets into developing solar, wind and nuclear projects. These companies are also supporting emerging technologies, such as Form Energy’s 100-hour batteries, through direct investments and working with utilities to accelerate adoption.
Dozens of startups are pursuing technologies that solve power problems. For example, Amperesand, DG Matrix, and Heron Power are developing new power conversion technologies, and companies such as Camus, GridBeyond, and Texture are building software that can manage the flow of electrons.
Power remains one of the most critical constraints for data centers, and this shortage is unlikely to change anytime soon. According to Goldman Sachs, AI is expected to increase data center power consumption by 175% by 2030.
Tech Crunch Event
San Francisco, California
|
October 13-15, 2026
These grid shortages are unprecedented in modern times and are causing electricity prices to rise across the country. This has led many technology companies to look for alternative ways to power their data centers. (The Trump administration, sensing a looming political crisis, is demanding tech companies build their own power sources, pay higher rates, or both. Of course, most have already made plans to do so.)
grid alternative
Amazon, Google, Oracle, and other large technology companies have been working to minimize their dependence on the grid. Several data centers are planned using on-site power or a hybrid approach that mixes on-site power with grid connections.
The largest data centers are driving this change. Less than a quarter of projects that have identified power sources will use on-site or hybrid. Together they account for 44% of total capacity.
These changes are due in part to a shortage of power generation equipment, such as gas turbines, and an outdated grid. This paves the way for alternative energy sources.
Google’s recent deal to build a new data center in Minnesota illustrates one approach to solving the problem. The company will combine Form Energy’s high-capacity 30-gigawatt-hour battery with wind and solar power. Google also worked with Xcel Energy to devise a new rate structure that will help encourage the adoption of new technologies in the utility planning process.
Form Energy’s batteries are not the only example. Grid-scale batteries are poised to make a big impact in the power market. By the end of this year, the United States will have about 65 gigawatts of battery storage capacity, according to the U.S. Energy Information Administration. Like many of its peers, Form Energy is looking to capitalize on the momentum by raising $500 million ahead of an eventual IPO.
An Underrated Skill
Energy supply is only part of the story. Once power is delivered to the grid or data center, it must be managed, and this is often the responsibility of the humble transformer.
Most transformers today use huge blocks of iron wrapped in copper wire, a technology that is about 140 years old. Although reliable, they are becoming too bulky as data centers’ power demands increase. One expert told TechCrunch that by the time the power density of server racks reaches 1 megawatt, the power equipment needed to run them will take up twice as much space as the racks themselves.
That’s why investors have recently flocked to back solid-state transformer startups that hope silicon-based power electronics can replace ancient iron and copper technologies. Although it is more expensive than traditional transformers, it is flexible enough to replace multiple pieces of equipment in a data center, making it cost-competitive.
Overall, the size of the investment in battery and transformer companies was much smaller than some of the blockbuster rounds we’ve seen in the AI industry.
That’s not a bad thing. These rounds are more manageable for investors. Additionally, as the world electrifies everything from transportation to heavy industry, electricity demand will continue to grow and investors may brace for an AI disruption. Perhaps the best AI investments may not be in AI at all.
