--- title: "AI Electricity Demand Expected to Triple Within Five Years; US Grid 'Unprepared' as Natural Gas Takes Center Stage" type: "News" locale: "en" url: "https://longbridge.com/en/news/284501583.md" description: "The rate at which AI data centers consume electricity has far exceeded the carrying capacity of the US grid. Peak demand is forecast to surge threefold within five years, making natural gas the only option capable of meeting both speed and scale requirements. However, these power plants, with operational lifespans of up to 30 years, are widening the gap between carbon reduction commitments and energy realities" datetime: "2026-04-29T06:00:44.000Z" locales: - [zh-CN](https://longbridge.com/zh-CN/news/284501583.md) - [en](https://longbridge.com/en/news/284501583.md) - [zh-HK](https://longbridge.com/zh-HK/news/284501583.md) --- # AI Electricity Demand Expected to Triple Within Five Years; US Grid 'Unprepared' as Natural Gas Takes Center Stage Artificial intelligence is pushing the US power grid to its design limits. According to Goldman Sachs Research, global data center electricity demand is projected to reach 84 gigawatts (GW) by 2027, a 50% surge from 2023, with AI workloads accounting for 27%. In the United States, utilities’ forecasts for summer peak demand over the next five years have jumped from 38 GW to 128 GW, more than doubling within a single planning cycle. **The implication behind these figures is that the US grid was already nearing its carrying capacity before the AI wave arrived.** Faced with this shortfall, natural gas has become the only option currently capable of meeting demand in terms of both speed and scale. Entergy is investing $3.2 billion to build three natural gas power plants with a combined capacity of 2.3 GW, specifically to power Meta’s new data center in Louisiana. NextEra Energy, the largest renewable energy developer in the US, is partnering with ExxonMobil to construct a 1.2 GW natural gas plant in the southeastern US. NextEra CEO John Ketchum summarized this trend by stating that the AI industry is shifting toward "BYOG"—Build Your Own Generation. However, the cost of betting heavily on natural gas is becoming apparent. The capacity market clearing price for the 2026–2027 delivery year in the PJM Interconnection (covering most of the Mid-Atlantic and Midwest regions) has surged to $329 per megawatt, more than ten times the $28.92 per megawatt recorded two years ago. The installation cost for new combined-cycle gas turbines has also nearly doubled to approximately $2,000 per kilowatt. Meanwhile, these plants, designed for an operational lifespan of about 30 years, will profoundly impact US carbon reduction commitments. ## Natural Gas Becomes the Only Viable Option Currently, about 40% of US electricity comes from natural gas, with renewables and coal making up the rest. AI data centers require stable, gigawatt-scale power around the clock, which is precisely the weak link in the existing grid. Renewables face hard constraints here: the median wait time for interconnection applications for new solar and wind projects exceeds four years. In contrast, natural gas is cheap and abundant, and leveraging the national pipeline network allows new plants to become operational within three to five years. The three plants Entergy is building for Meta’s Louisiana data center require 2 GW of power just for computing loads. The partnership plant between NextEra and ExxonMobil bets on the logic of "building infrastructure first, then attracting tenants." John Ketchum’s "BYOG" statement effectively marks a shift in the relationship between energy companies and tech giants from "supplier and consumer" to "joint infrastructure builders." ## Power Grid Engineering Faces Systemic Restructuring Traditional grid engineering is built on the premise of predictable loads—seasonal peaks, industrial cycles, and population growth can all be modeled and planned years in advance. AI has completely shattered this logic. **Large language model training tasks mean thousands of GPUs running at high load simultaneously for days or even weeks, followed by abrupt power drops; meanwhile, every AI inference response triggers instantaneous power spikes at the hundreds of megawatts level (inference spike). Such load behavior has no historical precedent, directly impacting the dispatch curves and reserve capacity models that utilities rely on.** To address this issue, gas peaker plants—designed for short-term high output—are being deployed directly adjacent to data center campuses to absorb inference pulses that base-load plants cannot match in response speed. Meanwhile, the Electric Reliability Council of Texas (ERCOT) has developed a new "Adjusted Large Load Forecast" methodology to distinguish actual realized demand from "phantom demand" created by speculative interconnection applications. ## Transmission Bottlenecks and Soaring Costs The physical grid is also under pressure. Transmission investment in many parts of the US has continued to shrink since 2015, leaving the system operating near its limits even before the AI demand surge. A national transmission needs study by the US Department of Energy (DOE) determined that transmission congestion in multiple regions across the country was already severe before this round of demand growth. Demand-side data is equally alarming: CenterPoint Energy in Texas reported a 700% increase in large-load interconnection applications between late 2023 and late 2024; Virginia currently has up to 50 GW of data center projects waiting in the interconnection queue. Market prices reflect this structural tightness: PJM capacity market clearing prices soared from $28.92 per megawatt to $329 per megawatt within two years; the installation cost for new combined-cycle gas turbines also nearly doubled from levels seen a few years ago to approximately $2,000 per kilowatt. These costs will ultimately be passed down through electricity pricing mechanisms. ## Natural Gas Lock-in Effect and Carbon Emission Risks The natural gas plants currently being approved and constructed are not merely transitional infrastructure. With an average operational lifespan of about 30 years, they will remain in operation well beyond the deadlines for almost all major net-zero emission targets. The lifecycle carbon emissions of natural gas are approximately 490 grams of CO₂ per kilowatt-hour. In the southern US, utilities plan to add about 20 GW of natural gas capacity over the next 15 years, with 65% to 85% of the projected load growth in Virginia, South Carolina, and Georgia coming from data centers. The methane issue further exacerbates carbon emission risks: natural gas infrastructure (drilling, pipelines, compression) continuously leaks methane, which has a warming effect approximately 80 times that of CO₂ over a 20-year timeframe. This issue is gradually evolving into a policy fault line among energy companies, large tech firms committed to net-zero emissions, and regulators who have yet to clarify the energy consumption impacts of AI. ## Policy and Alternative Paths: Distant Solutions Cannot Quench Immediate Thirst Several structural policy tools are advancing, but none can resolve the current supply-demand gap. Regarding energy storage, the Inflation Reduction Act (IRA) of 2022 provides a 30% tax credit for independent storage systems commissioned after 2024, covering the storage facilities themselves, thereby offering financial incentives for operators and utilities to invest in battery systems. Regarding nuclear power, as a zero-carbon option capable of providing stable base-load electricity, nuclear energy is receiving increasing attention. Google has reached an agreement with NextEra to restart the 615-megawatt Duane Arnold nuclear facility to secure round-the-clock carbon-free power. Transmission expansion remains the hardest nut to crack. DOE studies show that nearly every region in the US faces significant transmission capacity gaps, and bridging these gaps will require years of coordinated investment and permitting reform. Currently, the growth rate of AI technology demand has comprehensively outpaced the construction speed of the infrastructure it relies on—natural gas plants, transmission upgrades, energy storage deployment, and nuclear restarts have all failed to keep up. How this gap is bridged will depend on whether policy coordination and proactive investment can precede more painful consequences such as power shortages, data center construction delays, and rising electricity prices. ### Related Stocks - [159588.CN](https://longbridge.com/en/quote/159588.CN.md) - [FCG.US](https://longbridge.com/en/quote/FCG.US.md) - [XLE.US](https://longbridge.com/en/quote/XLE.US.md) - [UNG.US](https://longbridge.com/en/quote/UNG.US.md) - [BOIL.US](https://longbridge.com/en/quote/BOIL.US.md) - [GS.US](https://longbridge.com/en/quote/GS.US.md) - [ETR.US](https://longbridge.com/en/quote/ETR.US.md) - [META.US](https://longbridge.com/en/quote/META.US.md) - [NEE.US](https://longbridge.com/en/quote/NEE.US.md) - [XOM.US](https://longbridge.com/en/quote/XOM.US.md) - [CNP.US](https://longbridge.com/en/quote/CNP.US.md) - [GOOGL.US](https://longbridge.com/en/quote/GOOGL.US.md) - [GOOG.US](https://longbridge.com/en/quote/GOOG.US.md) - [W4VR.SG](https://longbridge.com/en/quote/W4VR.SG.md) - [NEE-N.US](https://longbridge.com/en/quote/NEE-N.US.md) - [NEE-W.US](https://longbridge.com/en/quote/NEE-W.US.md) - [NEP.US](https://longbridge.com/en/quote/NEP.US.md) - [NEE-R.US](https://longbridge.com/en/quote/NEE-R.US.md) - [NEE-T.US](https://longbridge.com/en/quote/NEE-T.US.md) - [NEE-U.US](https://longbridge.com/en/quote/NEE-U.US.md) - [NEE-S.US](https://longbridge.com/en/quote/NEE-S.US.md) - [NEE-V.US](https://longbridge.com/en/quote/NEE-V.US.md) ## Related News & Research - [Colder US Weather Forecasts Lift Nat-Gas Prices](https://longbridge.com/en/news/284260732.md) - [Front Month Nymex Natural Gas Rose 1.07% to Settle at $2.5500 — Data Talk](https://longbridge.com/en/news/284259976.md) - [Furbabies AI Expands Access to Pet Care and Reduces Loneliness](https://longbridge.com/en/news/284461616.md) - [SAS AI Navigator to bring order to AI chaos](https://longbridge.com/en/news/284419261.md) - [AI Trading Tools Are Getting Smarter — Here’s What Traders Need to Know](https://longbridge.com/en/news/283859061.md)