Analysis concluded US power grids cannot keep pace with datacenter power demand growth; new transmission and generation are structural bottlenecks.
Growth in artificial intelligence is outpacing the world's ability to generate enough power to sustain it. As AI adoption accelerates worldwide, it is colliding with a power system that was not built for its level of demand. According to Vittorio Pierangeli, senior vice president of PowerGen at Rolls-Royce Power Systems, electricity—not chips or algorithms—is the defining factor of the AI era.
The scale of this shift becomes clear by looking back less than a decade. In 2018, the AI-generated portrait "Edmond de Belamy" made global headlines as a novelty, trained on just 5 to 7 gigabytes of data before selling at Christie's for $432,500. Today, roughly 43 million AI-generated images are produced daily. The models behind them are trained on billions of data points—as much as 100,000 times the computational load of that early experiment. And image generation is only one slice of how AI is now woven into daily life.
The power intensity of AI operations far exceeds traditional computing. A single ChatGPT query can consume 10 to 100 times more energy than a standard Google search, depending on prompt complexity. On the data center floor, AI server racks now draw 100 to 120 kilowatts of power each, versus roughly 10 kilowatts just five years ago.
These pressures compound amid a changing energy mix and geopolitical uncertainty. Conventional coal plants are being decommissioned while renewable generation remains intermittent. Geopolitical instability is increasing pressure on energy security, and the load profiles of AI data centers are growing dramatically more volatile.
Projections suggest the entire power generation market will nearly triple between 2025 and 2030, driven largely by data center demand, with the backup power segment expanding roughly 22% annually and continuous power growing at 24% annually. Supply, however, is not keeping pace. Cumulative U.S. grid power supply to data centers is forecast to fall more than 50 gigawatts short of demand by 2030. While a data center itself can be built in 18 to 24 months, securing a grid connection takes three to seven years.
To bridge this gap, hyperscalers are increasingly turning to on-site power generation. Solutions such as mtu gas gensets provide near-term capacity, with small modular reactors offering longer-term alternatives. Developers are folding independent power plants into new data center designs under "bring-your-own-power" (BYOP) initiatives, generating the energy a facility needs until grid capacity or new nuclear sources come online. These systems offer short deployment times, high operational flexibility, modular design, and high efficiency—critical advantages in markets where natural gas is plentiful and pipeline infrastructure is well-established.
Perhaps the toughest challenge of the AI era is not the volume of power needed but how erratically it is needed. Graphics processing units operate synchronously, generating sharp, regular power swings. Research shows that within a single 50-megawatt block of an AI data center, real power can swing by ±20 megawatts within seconds. These fluctuations can strain transformers and disrupt the broader grid.
Kinetic energy storage systems such as mtu Kinetic PowerPacks act as fast-reacting buffers, stabilizing voltage and frequency while smoothing out power peaks. They respond instantly to load changes and provide uninterruptible power supply functionality without additional batteries.
Even with prime power secured through either a BYOP scheme or the grid, data centers require a failsafe backup plan. To achieve the Uptime Institute's Tier IV certification—requiring 99.99% availability—modern data centers depend on mission-critical diesel backup systems. More than 25% of data centers globally are backed up by mtu Series 4000 gensets.
Data centers are increasingly described as the invisible backbone of modern life—the factories of the digital age, where intelligence is produced and delivered. Powering them reliably and at the necessary pace requires solutions that can move as fast as the industry itself. Modular, scalable, and fuel-flexible systems meet data center operators where they are today and where they need to be tomorrow.