Nvidia, Utility Stocks Gain as AI Power Demand Puts Data-Center Electricity Use on Track for 2030 Highs

March 23, 2026
Nvidia, Utility Stocks Gain as AI Power Demand Puts Data-Center Electricity Use on Track for 2030 Highs

HOUSTON, March 23, 2026, 10:25 CDT

  • Nvidia and Emerald AI rolled out a “power-flexible” AI data center blueprint just as EPRI introduced its grid-flexibility framework at CERAWeek in Houston. 1
  • U.S. data centers might take up as much as 9%-17% of the nation’s electricity by 2030, EPRI said—well above the current 4%-5%. 2
  • Nvidia, Constellation, and Vistra shares each traded up during the U.S. morning session.

Nvidia and Emerald AI said Monday they’re teaming up with U.S. utilities—including Constellation, NextEra Energy, and Vistra—to develop “power-flexible” AI data centers. On the same day, EPRI introduced a voluntary framework aimed at trimming the time it takes these centers to lock in electricity. The announcements landed as CERAWeek kicked off in Houston, highlighting a challenge that’s suddenly front and center for the AI sector: power. 1

The load isn’t small anymore, and that’s significant. According to EPRI last month, U.S. data centers could soak up as much as 9% to 17% of nationwide electricity generation by 2030—up sharply from the current 4% to 5%. The Energy Information Administration is also projecting that U.S. power demand will hit new highs in both 2026 and 2027. If demand runs hotter than baseline forecasts, the agency’s higher-case scenario shows a much faster jump in gas-fired generation. 2

That shift is already getting priced in. Nvidia was up 3.1% at $178.13 this morning in the U.S., while Constellation picked up 4.2% to $293.92 and Vistra pushed 6.6% higher to $155.71.

EPRI’s Flex MOSAIC aims to speed up “time to power”—how quickly a project can get electricity online—by standardizing how utilities, regulators, and developers label a site’s load flexibility: how much power it can trim or shift, for how long, and how frequently. “Flexibility is becoming the third leg of the speed-to-power stool,” said Chief Executive Arshad Mansoor. 3

Nvidia’s DSX Flex software will give major AI facilities tools to shift power demand on the fly, syncing with both grid signals and their own batteries or generators. “Every system must be designed together,” CEO Jensen Huang said, referencing energy and cooling integration. Nvidia is targeting a commercial-scale launch at its AI Factory Research Center in Virginia later this year. 1

The impact won’t be spread out evenly. According to EPRI, Virginia—currently the nation’s largest data-center hub—could see the industry’s electricity usage jump from around 25% now to somewhere between 39% and 57% by 2030. In Arizona, Indiana, Iowa, Nebraska, Nevada, Oregon and Wyoming, projections show the sector accounting for more than 20% of total demand in each state. 2

Big Tech companies are already experimenting with ways to flatten those demand spikes. Just last week, Google announced it had reached five agreements with utilities, giving it the ability to trim up to 1 gigawatt from its data centers during periods of severe grid strain. This approach, demand response, involves major power users scaling back when the grid is under pressure. Michael Terrell, who heads advanced energy at Google, described the strategy as “a really important tool” for handling what’s coming. 4

Some players are opting for dedicated supply instead. On Friday, the U.S. Department of Energy announced SB Energy, part of SoftBank, and AEP are behind plans for a 10-gigawatt AI campus alongside a 9.2-gigawatt gas plant on federal land in Ohio. It’s a move that signals certain developers prefer to sidestep grid congestion by building their own solutions. 5

It’s a global story. The International Energy Agency is forecasting that data center power usage will climb to roughly 945 terawatt-hours by 2030—a standard unit for electricity consumption—while electricity generated for these facilities is set to top 1,000 TWh. Renewables are on track to supply close to half of that additional demand; even so, gas and coal combined are anticipated to account for more than 40% of the growth through 2030. 6

Still, fixes are moving slowly. EPRI’s framework is strictly voluntary. NARUC President Ann Rendahl notes that regulators remain intent on shielding households from footing the bill for new, large power users. Last month, Reuters flagged PJM’s warning: possible supply shortfalls by 2027. Meanwhile, big turbine makers are booked solid into the late 2020s. 3

Joe Dominguez, CEO of Constellation, pushed back on the notion that it’s just about overall supply. “We don’t have a supply problem — we have a peak problem,” he said. 1

Technology News

  • GeForce Now adds 90 FPS VR streaming for Quest, Vision Pro, Pico headsets
    March 23, 2026, 11:50 AM EDT. NVIDIA's GeForce Now adds a 90 FPS upgrade for VR streaming to Ultimate subscribers, expanding support to Meta Quest, Apple Vision Pro, and Pico headsets. The jump from 60 FPS aims at crisper motion that matches headsets' native refresh rates. Separately, Vision Pro users can access 4K 120 FPS via Nvidia's CloudXR, a stream that runs outside GeForce Now. Nvidia calls cloud VR a way to transform the space into a personal gaming theater rather than a full VR experience with motion controls. The rollout begins today, timed with Crimson Desert's GeForce Now launch and its RTX 5080-class servers for smoother performance.

Latest Articles

Nvidia, Utility Stocks Gain as AI Power Demand Puts Data-Center Electricity Use on Track for 2030 Highs

Nvidia, Utility Stocks Gain as AI Power Demand Puts Data-Center Electricity Use on Track for 2030 Highs

March 23, 2026
Nvidia and Emerald AI announced a partnership with U.S. utilities to develop “power-flexible” AI data centers, as EPRI launched a framework to speed up electricity access for such projects at CERAWeek in Houston. EPRI warned U.S. data centers could use 9%-17% of national electricity by 2030, up from 4%-5% now. Nvidia, Constellation, and Vistra shares traded higher Monday morning.
China’s Open-Source AI Surge Challenges U.S. Tech Dominance as Alibaba Stock Gains

China’s Open-Source AI Surge Challenges U.S. Tech Dominance as Alibaba Stock Gains

March 23, 2026
A U.S. advisory commission warned Monday that Chinese open-source AI models now dominate usage on platforms like Hugging Face, outpacing U.S. counterparts. Alibaba launched Accio Work for small businesses, and Tencent integrated an AI agent into WeChat. Siemens’ CEO said there are “no disadvantages” to using Chinese open-source AI for industrial automation. Chinese models are now estimated to be used in about 80% of U.S. AI startups.