On Aug 2 the South Korean city of Yangsan hit 42.5 degrees Celsius — the highest temperature in 122 years of record-keeping. Seoul received its first-ever severe heat wave warning. More than 20 people have died, most of them over 80.
On Aug 4, President Lee Jae-myung called the heat wave a national disaster. Then he gave an order that tells you everything about the modern world: inspect the power system.
He was right to. In a heat wave, the grid is the difference between discomfort and catastrophe. Air conditioning, hospital cooling, refrigerated medicine, water pumping — all of it hangs on wires most people never think about until the lights go out.
Europe is counting its heat dead again this summer. Record electricity demand is being set across the Northern Hemisphere. Behind every one of those headlines sits a question ordinary readers understand before they finish reading it: can the grid withstand a hotter world?
I have spent my career on that question. In January 1998, at the Electric Power Research Institute in California, I created and led a research program built on an unusual premise that the power grid could learn to heal itself. We called it the Complex Interactive Networks/Systems Initiative — 108 professors at 28 universities, 52 utilities and the US Department of Defense. Out of that work came the concept, and the name, of the smart grid: a network with sensors for nerves, computation for reflexes and the ability to isolate damage before it spreads.
Here is what heat does to that network, in plain terms. Demand surges as millions of air conditioners switch on at once. At the same moment, the grid itself weakens. Transmission lines sag and carry less power. Transformers run hot and age faster.
Power plants lose efficiency in warm air and warm cooling water. The system is asked to do the most precisely when it can do the least. A smart grid exists for this moment — to see the stress building, shift the load and keep the most important circuits alive.
Twenty-eight years after those whiteboard sketches in Palo Alto, the largest laboratory for these ideas is on the other side of the Pacific. China crossed 10,000 terawatt-hours of electricity consumption in 2025 — the first nation ever to do so.
State Grid Corporation of China invested a record 650 billion yuan ($96.33 billion) last year.
The new five-year plan calls for more than 5 trillion yuan of grid investment through 2030, including 15 new ultra-high-voltage corridors carrying wind and solar power from the western deserts to the eastern coast.
The United States and Europe are taking a slower, market-driven path. The strain shows. The International Energy Agency counts more than 2,500 gigawatts of projects stalled in grid-connection queues worldwide. Global grid investment runs about $400 billion a year; the IEA says it must grow by half again by 2030 just to keep pace.
Generation has raced ahead. The wires have not.
And the heat is only the first wave. The second is already arriving: artificial intelligence. Data centers consumed roughly 485 terawatt-hours globally in 2025 and are on track to approach 950 by 2030. Electricity demand from AI-focused facilities alone grew 50 percent last year. The two pressures now land on the same wires, in the same summers.
Three great experiments are running at once. China is testing whether centralized planning can build transmission fast enough to carry a renewable buildout of unprecedented scale. The US is testing whether markets can reprice reliability quickly enough to serve an AI boom nobody scheduled. Europe is testing whether interconnected national grids can hold together under stress — a question made vivid by the April 2025 Iberian blackout, when an overvoltage-driven cascade darkened two countries in minutes.
Each experiment will teach the others. None can afford to fail alone.
What should we do? The engineering answers are known and unglamorous. Build transmission before it is needed, not after. Put sensing and automated protection deep into distribution networks, where most outages begin. Treat demand flexibility — the ability of data centers, factories and buildings to shift their load — as a resource equal to any power plant. Train the workforce; the IEA warns that people and supply chains, not technology, are the tightest constraints. And share what we learn across borders, because heat ignores them, and electrons obey physics, not politics.
It also means remembering what a grid is for. A grid is not a market abstraction or a policy instrument. It is the quiet machine beneath every hospital ventilator, every water pump, every fan turning in an elderly person's room in Yangsan or Yeongdeungpo in South Korea. When engineers argue about transmission capacity and reserve margins, the real subject is whether ordinary people can trust the wall socket.
Repair is a decision. I trust it because I have watched it happen, in power systems and elsewhere: the deliberate, unheroic choice to strengthen what holds rather than argue about what failed.
Laozi (Lao Tzu) wrote: "Do the difficult things while they are easy, and do the great things while they are small." The grid taught me the same lesson in a different language. The time to strengthen the foundation is before the weight arrives.
This summer, the weight arrived.
On Aug 2 the South Korean city of Yangsan hit 42.5 degrees Celsius — the highest temperature in 122 years of record-keeping. Seoul received its first-ever severe heat wave warning. More than 20 people have died, most of them over 80.
On Aug 4, President Lee Jae-myung called the heat wave a national disaster. Then he gave an order that tells you everything about the modern world: inspect the power system.
He was right to. In a heat wave, the grid is the difference between discomfort and catastrophe. Air conditioning, hospital cooling, refrigerated medicine, water pumping — all of it hangs on wires most people never think about until the lights go out.
Europe is counting its heat dead again this summer. Record electricity demand is being set across the Northern Hemisphere. Behind every one of those headlines sits a question ordinary readers understand before they finish reading it: can the grid withstand a hotter world?
I have spent my career on that question. In January 1998, at the Electric Power Research Institute in California, I created and led a research program built on an unusual premise that the power grid could learn to heal itself. We called it the Complex Interactive Networks/Systems Initiative — 108 professors at 28 universities, 52 utilities and the US Department of Defense. Out of that work came the concept, and the name, of the smart grid: a network with sensors for nerves, computation for reflexes and the ability to isolate damage before it spreads.
Here is what heat does to that network, in plain terms. Demand surges as millions of air conditioners switch on at once. At the same moment, the grid itself weakens. Transmission lines sag and carry less power. Transformers run hot and age faster.
Power plants lose efficiency in warm air and warm cooling water. The system is asked to do the most precisely when it can do the least. A smart grid exists for this moment — to see the stress building, shift the load and keep the most important circuits alive.
Twenty-eight years after those whiteboard sketches in Palo Alto, the largest laboratory for these ideas is on the other side of the Pacific. China crossed 10,000 terawatt-hours of electricity consumption in 2025 — the first nation ever to do so.
State Grid Corporation of China invested a record 650 billion yuan ($96.33 billion) last year.
The new five-year plan calls for more than 5 trillion yuan of grid investment through 2030, including 15 new ultra-high-voltage corridors carrying wind and solar power from the western deserts to the eastern coast.
The United States and Europe are taking a slower, market-driven path. The strain shows. The International Energy Agency counts more than 2,500 gigawatts of projects stalled in grid-connection queues worldwide. Global grid investment runs about $400 billion a year; the IEA says it must grow by half again by 2030 just to keep pace.
Generation has raced ahead. The wires have not.
And the heat is only the first wave. The second is already arriving: artificial intelligence. Data centers consumed roughly 485 terawatt-hours globally in 2025 and are on track to approach 950 by 2030. Electricity demand from AI-focused facilities alone grew 50 percent last year. The two pressures now land on the same wires, in the same summers.
Three great experiments are running at once. China is testing whether centralized planning can build transmission fast enough to carry a renewable buildout of unprecedented scale. The US is testing whether markets can reprice reliability quickly enough to serve an AI boom nobody scheduled. Europe is testing whether interconnected national grids can hold together under stress — a question made vivid by the April 2025 Iberian blackout, when an overvoltage-driven cascade darkened two countries in minutes.
Each experiment will teach the others. None can afford to fail alone.
What should we do? The engineering answers are known and unglamorous. Build transmission before it is needed, not after. Put sensing and automated protection deep into distribution networks, where most outages begin. Treat demand flexibility — the ability of data centers, factories and buildings to shift their load — as a resource equal to any power plant. Train the workforce; the IEA warns that people and supply chains, not technology, are the tightest constraints. And share what we learn across borders, because heat ignores them, and electrons obey physics, not politics.
It also means remembering what a grid is for. A grid is not a market abstraction or a policy instrument. It is the quiet machine beneath every hospital ventilator, every water pump, every fan turning in an elderly person's room in Yangsan or Yeongdeungpo in South Korea. When engineers argue about transmission capacity and reserve margins, the real subject is whether ordinary people can trust the wall socket.
Repair is a decision. I trust it because I have watched it happen, in power systems and elsewhere: the deliberate, unheroic choice to strengthen what holds rather than argue about what failed.
Laozi (Lao Tzu) wrote: "Do the difficult things while they are easy, and do the great things while they are small." The grid taught me the same lesson in a different language. The time to strengthen the foundation is before the weight arrives.
This summer, the weight arrived.
