A standard nuclear power plant generates a gigawatt or more of low-carbon power, a boon in this age of anxiety over climate change. The problem is getting the thing built in the first place: At US $7 billion to $10 billion apiece, nuclear plants are tough for even the largest utilities to finance.

President Obama proposes to handle the problem by tripling federal loan guarantees to such plants, to $54 billion. But now a more economical solution is coming under scrutiny: downsizing nuclear plants from gigawatt scale to more affordable units that can be built by the dozen. ”Size matters. In this case, small size,” says Andrew Kadak, a professor of nuclear science and engineering at MIT.

Small modular reactors, or SMRs, of 70 to 210 megawatts are under construction in China and Russia, and a mix of start-ups and established nuclear technology firms, such as Westinghouse Electric Co., General Atomics, and the Babcock & Wilcox Co., are shopping similarly modest designs in the United States.

This strategy overturns the drive toward economies of scale that has pushed nuclear designers toward ever-larger reactors since the industry’s inception. Now the designers may instead rely on the ”economies of multiples” that accrue to the mass production of everything from cars to iPhones.

”We want to manufacture in a plant with supply-chain management. This enables you to drive down cost and control the schedule,” says John Parmentola, senior vice president for energy and electromagnetic systems at General Atomics. That means building modules, including reactors, that are small enough to be shipped on a truck or railcar and designed so that they can be snapped together on-site. ”It’s almost Lego-style assembly,” says Kadak.

These innovators hope to avoid the sprawling construction sites required to build today’s gigawatt-plus reactors, which are prone to quality problems and delays. For example, in 2005 France’s Areva boasted that its flagship 1.65-gigawatt pressurized water reactor, the EPR, would be completed by 2009. Now the company is admitting that faulty materials and planning snafus have set the completion target back to 2012 and raised the project’s estimated cost by 66 percent, to a budget of 5.3 billion ($7.2 billion).

Proponents of SMRs admit that their installation costs may turn out to be as much as or even more than that of today’s behemoths, but they argue that the lower risk involved should make SMRs the better deal anyway. Christofer Mowry, CEO of Babcock & Wilcox’s Modular Nuclear Energy subsidiary, leads the development of a 125-MW SMR called mPower that he estimates will cost about $600 million in parts and labor. That’s comparable to Areva’s Olkiluoto plant on a per-megawatt basis, but because mPower could be built in bite-size chunks with a relatively modest overhead investment, using the same reliable light-water reactor technology, it’s much more likely to work and to start working on schedule. That means the cost of financing and insuring the project should be much lower. ”You could have 10 to 20 percent cheaper electricity,” says Mowry.