Gellings: Load is growing about twice as fast as transmission capacity, and has been for over 10 years. Lots of congestion. We've modernized virtually every industry in the U.S. except this one -- it's mechanically controlled, no sensor, no information technology, no digitization. It can't heal itself. We get info about problems too late. And that's just the beginning.

Ambrosio: Utilities are the last to digitize. IBM is eager to take on this computational problem.

Pullins: At the transmission level, on 60% of the network has SCATA, and only 2% of distribution. The grid is operating at the speed of light, but mechanically, with no sophisticated control.

Gellings: Only 7.5% of feeders have tight voltage control. We could save 1% of the nation's electricity by improving that.

Burns: The reason we want to improve all these statistics is for consumers. They need more reliability and better price mechanisms.

What's changed that grid improvements are possible now?

Gellings: Don't things work relatively well? No. The problems are hidden but pervasive. Why spend all the money to improve it? This stuff costs consumers $180 billion a year, but those costs are diffused and hidden. Just now we're seeing dialogue between utilities and regulators about improvements. It would cost about $600 billion to fix it.

Pullins: No, it would cost $0. We already have plans in place to spend over $900 billion on this stuff. We just need to redirect that money.

What is a smart grid?

Pullins: We (modern grid initiative) began in 2004; started with systems analysis. We found seven smart grid characteristics:

• motivate and incorporate consumer: a) utilize assets consumer has, mainly load assets, occasionally generation assets, b) engage consumers in conservation and efficiency efforts
• accommodate wide variety of generation and storage
• accommodate competitive markets
• resist attack
• match power quality to needs
• optimize assets
• self-healing

All the true innovation in the grid is on the margins, with consumer electronics and such, not happening from federal or utility research dollars. We're at a point where our infrastructure is becoming less and less relevant to the electric service the consumer base expects to have.

Ambrosio: Yes, things are happening at the edge, out in operations. There's an analogy with the computer industry. There was a thought in the '80s, '90s that everything was going to be distributed to the desktop. But no: high-performance central computing stayed around, it just became much more sophisticated and integrated with distributed computation. The model changed.

So, we need more processing and analysis in the poles, on the edge, just like distributed programming.

Gellings: Explaining "Galvin." He wanted perfect power that would not fail. People studied it. And ... he completely lost me.

What's the future of the interface between consumers and utilities? (Currently, it's a dumb, electromechanical meter.)

Burns: Smart metering is just about knowing more about what your customers want and need. Smart metering will allow utilities to measure consumption at an almost minute-by-minute basis. You start to understand what customers are really doing.

Smart meters not only communicate back and forth to utilities, they extend into the home, providing information to consumers about their real-time consumption. Smart meters mean smart consumers.

How do you lower emissions?

Gellings: Get more efficient. You need at least 7GW of renewables. You need nuclear and clean coal with CCS. You need plug-in hybrids.

Pullins: Right now we've got single-digit penetration of renewables. Germany, Austria, and Spain are about 13%, with a traditional grid. Denmark, with a smarter microgrid approach, is up to 32%. So that's how you enable renewables penetration: smart grid. You also get energy security, you get lower rates, and you get tons of unanticipated benefits.