So much sunlight hits the Earth each day that the world’s entire electricity needs could be met by harvesting only 2% of the solar energy in the Sahara Desert. Of course, using solar power as the world’s only energy source hasn’t been possible yet, in part because solar equipment is expensive to make (and getting the power out of the desert would be no easy feat, either). But researchers at IBM think they’re one step closer to making solar universally accessible with a low-cost system that can concentrate the sunlight by 2,000 times.

The system uses a dish covered in mirrors to aim sunlight in a small area; as the sun moves throughout the day, the dish follows it to catch the most light. Other concentrated solar power systems do the same thing, but quirks of this design make it much more efficient: A typical system only converts around 20% of the incoming light to usable energy, while this one can convert 80%.

That makes it cheap. Bruno Michel, manager for advanced thermal packaging at IBM Research, believes the design could be three-times cheaper than "comparable" systems. “The levelized cost of energy will be less than 10 cents per kilowatt hour," he says.

Using simple materials also helps. “The reflective material we use for the mirror facets are similar to that of potato chip bags,” Michel says. “The reinforced concrete is also similar to what is being used to build bridges around the world. So outside of the receiver, which contains the photovoltaic chips, we are using standard materials.”

A few small high-tech parts will be built in Switzerland, where the prototype is currently being produced. But the main parts of the equipment could easily be built locally, wherever it's being used.

It's especially well-suited for sunny areas that happen to be dry. As the system runs, it can use excess heat that would normally be wasted to desalinate water. A large installation could provide clean drinking water for an entire town.

IBM researchers are working with partners at Airlight Energy, ETH-Zurich, and Interstate University of Applied Sciences Buchs NTB to finish building a large prototype, which will be ready by the end of the summer. After testing, they hope to start production at scale within 18 months.