AivexaNewsSearch
AI news for builders and product teamsChecked every hour

The data center boom has an upside: innovation for the US grid

Collected Oct 27, 2026

Speaking at TechCrunch Disrupt earlier this month, three energy-industry figures argued that the enormous power demand created by the AI buildout is doing more than stressing America's aging grid: it is also opening a market for modernization. The panelists were Apoorv Bhargava, founder and CEO of WeaveGrid; Tim Heidel, CEO of Veir; and Kristin Eberhard, vice president of policy at Rewiring America. Their framing was optimistic, though they acknowledged the strain. Concerns raised around the boom include higher electricity bills in the near term and, over a longer horizon, possible outages and other technical problems.

Bhargava described a grid dealing with many new variables at once. Renewable generation has added variability on the supply side, while electric vehicles, data centers and heat pumps have added substantial new load on the demand side. What utilities are now grappling with, in his account, is the ability to manage and orchestrate all of that. WeaveGrid's product is software for exactly that job: helping organizations deploy energy efficiently and optimizing connections in ways the company says reduce bills for utility customers.

Heidel's company takes the physical route. Veir sells superconductive power lines, which the company says move power far more efficiently than conventional cables. Its stated figures are that the lines can carry power five times farther than traditional cables while requiring 20 times less space. Veir raised $110 million in a Series C round earlier this month, capital that presumably funds building out and expanding grid infrastructure rather than just software layers on top of it.

For developers and product teams, the superconductive line claims are worth reading carefully. Superconductors carry current with zero resistance, but only below a critical temperature, which in practice requires cryogenic cooling along the line. That cooling equipment is a large part of why superconducting links have historically been confined to short, dense urban runs and demonstration projects. Veir's pitch appears to rest on reducing the footprint and extending the reach enough to make the approach viable at grid scale; the five-times-distance and twenty-times-space figures are the company's own, not independent measurements, and the economics will depend on installation and operating costs that weren't discussed on stage. In practice, transmission projects also face permitting timelines that software cannot compress.

Heidel tied the innovation directly to AI spending. He argued that the scale of investment in AI infrastructure is driving a great deal of innovation, with many new technologies moving off the shelf and beginning to reach scale, and that combined they can help make data centers good community neighbors. He added that the priority should be adopting those technologies as fast as possible to ease the next stage of growth.

Eberhard, whose organization advocates for household electrification, said the surge in data center demand has added urgency to modernization while also creating a climate in which industries are willing to try new things. She characterized the current period as one of improvisation, with a lot of "yes-and-ing": yes to more transmission lines, yes to more centralized power plants, and also yes to taking advantage of options that were previously left on the table when AI demand had not created the same urgency.

A useful way to read the panel is as a map of where the money and engineering effort are heading. One layer is software: orchestration and load management that extracts more capacity from the grid that already exists, which is comparatively fast to deploy and cheap to scale. Another is hardware: new conductors and lines that increase how much power can move through constrained corridors. A third is demand-side policy: electrifying homes and shifting when devices draw power. Each addresses a different bottleneck, and none alone resolves the mismatch between where power is generated and where data centers are being built.

A likely trade-off is that the fastest wins come from coordination rather than construction. Software that shifts EV charging and building load can relieve peaks without new wires, but it depends on utilities, device makers and customers sharing control and data. Transmission upgrades promise far more capacity but arrive on multi-year timelines. That asymmetry suggests near-term relief will lean on orchestration while hardware and policy investments mature in the background.

Why it matters: The AI data center buildout is becoming a grid-planning problem as much as a compute problem, and startups are positioning software, cabling and electrification policy as the answers. Developers and product teams working near energy infrastructure, EV charging or utility integrations should expect more demand for load-shifting and orchestration capabilities, and more scrutiny of where and when large facilities draw power. Whether the claimed efficiency gains from new conductors reach commercial scale is, by inference, the open variable that determines how much of the crunch is solved by building rather than by managing demand.

Read at TechCrunch · AI

Based on reporting from the original publisher. Visit the source for full context and later updates.

Publisher excerpt

As the AI boom stresses the existing power grid, all sorts of new opportunities for grid modernization are also emerging as a result, industry insiders say.