Venture capital firms have committed progressively larger sums to nascent AI companies, channeling more than five hundred billion dollars into this domain during the past half-decade.
However, presently, the most astute allocation of capital within AI could lie in the energy sector, as per an analysis from Sightline Climate. Investigators discovered that as much as half of all publicized data center initiatives could face postponements. A primary factor contributing to this is the availability of electricity.
From the 190 gigawatts of data center capacity the firm monitors, a mere 5 gigawatts are currently being built. Approximately 6 gigawatts of data center endeavors cataloged in Sightline’s records became operational last year. A significantly greater proportion — roughly 36% — experienced schedule setbacks extending into 2025. These holdups could eventually have repercussions for major corporations and other entities leveraging AI in their operations.
This imbalance between supply and demand presents a lucrative prospect for financiers. The rationale is as follows.
Major technology corporations such as Google and Meta have allocated substantial portions of their financial resources to advance solar, wind, and nuclear energy initiatives. Furthermore, these firms are backing innovative technologies, for example, Form Energy’s 100-hour battery, via direct capital injections and by collaborating with utility providers to hasten their implementation.
Numerous fledgling companies are researching solutions aimed at resolving the energy challenge. For example, Amperesand, DG Matrix, and Heron Power are engineering novel power transformation systems, whereas enterprises such as Camus, GridBeyond, and Texture are crafting software capable of overseeing the movement of electrical current.
Electricity continues to be among the most critical limitations for data centers, a deficiency unlikely to abate in the near future. AI is projected to boost data center energy usage by 175% by the year 2030, as per projections from Goldman Sachs.
TechCrunch gathering
San Francisco, CA
|
October 13-15, 2026
These deficiencies within the power grid are unparalleled in contemporary history, leading to an escalation in electricity costs nationwide. This situation has prompted numerous technology firms to investigate alternative methods for supplying energy to their data centers. (The Trump administration, foreseeing an impending political dilemma, is advocating for tech companies to establish their own power generation, incur elevated charges, or a combination thereof. Most had, naturally, already devised strategies for this.)
Alternatives to the Main Grid
Amazon, Google, Oracle, and other major technology corporations have been striving to lessen their reliance on the central power network. Multiple data centers are being conceptualized employing either localized power generation or a mixed strategy combining on-site energy with a grid link.
The largest data centers are spearheading this initiative. Fewer than twenty-five percent of ventures that have pinpointed an energy supply will utilize on-site or hybrid solutions; collectively, these account for 44% of overall capacity.
This transition has been partially propelled by scarcities of energy production apparatus — particularly gas turbines — and an outdated electrical network. This circumstance has paved the way for unconventional power origins.
Google’s most recent agreement to energize a novel data center in Minnesota illustrates a method for addressing this challenge. The corporation intends to combine wind and solar power with an enormous 30 gigawatt-hour battery from Form Energy. Google further collaborated with Xcel Energy to formulate an updated tariff system which, it claims, will foster the integration of new technologies into the utility’s strategic planning.
Form Energy’s battery is not the sole illustration. Large-scale grid batteries are prepared to capture a substantial share of the energy market. By the close of this year, the United States is anticipated to possess almost 65 gigawatts of battery storage capability, as reported by the U.S. Energy Information Administration. Akin to numerous competitors, Form Energy aims to leverage this impetus by securing a $500 million funding round preceding a prospective initial public offering.
Underappreciated Technology
Energy provision constitutes merely a segment of the narrative. Once electricity reaches the grid or the data center, its regulation becomes necessary, a responsibility primarily entrusted to the unassuming transformer.
The majority of modern transformers employ substantial iron cores encased in copper wiring, a method dating back approximately 140 years. While dependable, it is proving excessively cumbersome as the energy requirements of data centers escalate. Should server racks attain a power density of 1 megawatt, the requisite power apparatus to operate them will consume double the physical area of the rack itself, an authority informed TechCrunch.
This explains why investors have recently been converging to support solid-state transformer nascent companies, who anticipate that silicon-based power electronics can supersede the antiquated iron-and-copper methodology. These are pricier than current transformers, yet they possess sufficient adaptability to substitute multiple components within a data center, which ought to render them economically viable.
In totality, the magnitude of capital injected into battery and transformer firms has been considerably less significant than certain monumental funding rounds witnessed within the AI sector.
This is not necessarily disadvantageous — such funding cycles are more manageable for investors. Furthermore, as global electrification expands across various domains, from transit to manufacturing, the demand for energy is set to invariably climb, offering financiers a safeguard against a potential AI downturn. Perhaps the most advantageous AI-related investment might not reside within AI itself.
{content}
Source: {feed_title}