“No dropouts, no brownouts,” another tech called.

Electricity in the grid is like water coming into a house from a single main pipe and flowing out through many open faucets. When one is closed, the pressure in the others increases. Electricity’s the same, though it moves a lot more quickly than water-nearly 700 million miles an hour. And because New York City demanded a lot of power, the voltages-the electrical equivalent of water pressure-in the substations doing the extra work were running high.

But the system was built to handle this and the voltage indicators were still in the green.

What was troubling the supervisor, though, was why the circuit breakers in MH-12 had separated in the first place. The most common reason for a substation’s breakers to pop is either a short circuit or unusually high demand at peak times-early morning, both rush hours and early evening, or when the temperature soars and greedy air conditioners demand their juice.

None of those was the case at 11:20:20:003 a.m. on this comfortable April day.

“Get a troubleman over to MH-Twelve. Could be a bum cable. Or a short in the-”

Just then a second red light began to flash.

Critical failure.

NJ-18 offline.

Another area substation, located near Paramus, New Jersey, had gone down. It was one of those taking up the slack in Manhattan-12’s absence.

The supervisor made a sound, half laugh, half cough. A perplexed frown screwed into his face. “What the hell’s going on? The load’s within tolerances.”

“Sensors and indicators all functioning,” one technician called.

“SCADA problem?” the supervisor called. Algonquin’s power empire was overseen by a sophisticated Supervisory Control and Data Acquisition program, running on huge Unix computers. The legendary 2003 Northeast Blackout, the largest ever in North America, was caused in part by a series of computer software errors. Today’s systems wouldn’t let that disaster happen again but that wasn’t to say a different computer screwup couldn’t occur.