Electricity, not ideas, is the constraint

The AI Energy Bottleneck Calculator

At TED, Eric Schmidt argued the hidden limit on artificial intelligence is not model architecture or chips. It is raw electric power. Explore what the gap actually looks like.

"There's a real limit in energy. We need another 90 gigawatts of power in America. 90 gigawatts is 90 nuclear power plants in America. Not happening. We're building zero."Eric Schmidt, former Google CEO, speaking at TED

How much new AI power demand?

90 GW
90large nuclear plants needed (1 GW each)
0US homes' worth of average demand
0H100-class GPUs powered continuously
0square miles of solar farms (with storage)
0terawatt-hours per year at full load
0share of total current US generating capacity

The construction gap

Visualizing the reactors we are not building

Each square is one gigawatt-scale nuclear plant. Amber squares are plants your scenario requires. The tiny green count is large reactors currently under construction in the United States: effectively zero.

Why electricity is the bottleneck

Training and serving frontier AI models requires enormous, always-on electric load. A single hyperscale AI campus can draw as much power as a mid-size city, and demand is compounding: US data centers already consume roughly 4-5 percent of national electricity, with credible forecasts of 9-12 percent by 2030.

Schmidt's framing is blunt arithmetic. Ninety gigawatts of new firm capacity is about the output of ninety large nuclear reactors. The United States commissioned its most recent new reactors, Vogtle 3 and 4, after roughly 15 years and over 30 billion dollars, and currently has essentially no large reactors under construction. Grid interconnection queues stretch four to seven years, and gas turbine backlogs run to 2029.

The result: the constraint on AI progress may shift from algorithms and silicon to substations, transmission lines, and generation. Whoever solves power first, whether through nuclear restarts, small modular reactors, gas, or solar-plus-storage, gains a structural advantage in the AI race.

Assumptions used in this calculator

1 large nuclear plant = 1 GW. Average US home demand = 1.2 kW (about 10,500 kWh per year). One H100-class GPU with cooling and overhead = 1.5 kW. Solar with storage delivering firm power = about 10 acres per firm MW. US total generating capacity = about 1,300 GW. Figures are order-of-magnitude estimates for intuition, not engineering studies.

Power demand: unit conversions and fixed assumptions

Read the explanation

This saved calculator takes a continuous power scenario in gigawatts. Ninety gigawatts sustained for eight thousand seven hundred sixty hours gives seven hundred eighty-eight point four terawatt-hours in a year; the interface rounds that to seven hundred eighty-eight. Power and energy have different units, so the bars explicitly use separate scales. The source divides gigawatts by a fixed thirteen-hundred-gigawatt denominator for its percentage display, giving about six point nine percent. That denominator and continuous operation are authored assumptions, not verified present-day capacity or a forecast. The video explains the program arithmetic without adopting its stored claims about current infrastructure. The program assigns twelve hundred watts to each home equivalent and fifteen hundred watts to each GPU equivalent. Dividing ninety billion watts gives seventy-five million home equivalents or sixty million chip equivalents. Bars use six pixels per million equivalents, while the labels keep the different kinds of load visible. These counts are arithmetic comparisons, not measured household demand, deployed GPU inventory or a claim that every chip draws this power continuously. Changing the demand input scales both counts linearly. A one-gigawatt scenario similarly gives about eight hundred thirty-three thousand homes or six hundred sixty-seven thousand chips under those fixed assumptions. The source treats each gigawatt as one plant unit. Its construction constant is literally zero, so the unmet gap is one hundred percent for every positive slider value. Bars compare five and ninety assumed one-gigawatt units at six pixels each. This fixed value does not establish the actual number of reactors under construction or the feasibility of future grid projects. The solar comparison multiplies gigawatts by one thousand megawatts, ten acres per megawatt, and divides by six hundred forty acres per square mile; ninety gives about fourteen hundred six square miles. These fixed conversion assumptions omit capacity factor, storage, transmission, location and reliability. No real power-system forecast is run.

Super generates helpful tools and automates fact-checking across the internet proactively. If you enjoyed this tool, build your own with Super and share it with a friend.