The U.S. Advanced Research Projects Agency for Energy, which began its fifth annual summit today in Washington, D.C., is meant to kick-start early-stage technologies that could transform the energy industry, replacing fossil fuels and reducing dependence on foreign energy sources. But five years after ARPA-E was first funded, is it living up to the objectives (see “What ARPA-E Can’t Do” and “What ARPA-E Does Well: Making Connections”)?

Some experts say that political pressures are making it difficult for the agency to support the risky technologies it was created to fund. ARPA-E has had bipartisan support, but it has always struggled to get funding. Last year, the House voted to give the agency just 20 percent of what it had asked for in its budget request. As it turned out, the agency got much more than that: the omnibus spending bill that passed earlier this year set aside $275 million for the agency, about 75 percent of the budget request.

Part of the agency’s challenge is that it’s feeling pressure from two sides. On the one hand, it needs to show results to justify continued funding. That’s led it to fund some projects that have a higher chance of paying off, and to taking “fewer risks than they should have,” says David Victor, director of the Laboratory on International Law and Regulation at the University of California at San Diego. But playing it safe can also bring criticism from Congress. The chief complaint from a 2012 House hearing on ARPA-E was that it was funding some projects that could have been funded by companies instead of the government.

It’s a tricky balance. “If everything worked right away, I wouldn’t be taking big enough swings,” says Cheryl Martin, ARPA-E’s acting director. But, as she tells members of Congress, “if nothing ever worked, you should stop funding us.”

To survive, Martin says, it’s important that some of the agency’s projects pay off unexpectedly fast, delivering tangible results. For example, one of the highlights of the summit’s technology showcase this year is a Toyota that’s been retrofitted with a new kind of battery charger invented by researchers at the University of Arkansas and Toyota using a $4 million ARPA-E grant—the charger is far smaller than conventional ones, and could charge batteries more quickly as well. She also points to projects that, while still a long way from market, have made enough progress that others have decided to invest in them.

Meanwhile, to keep finding areas where no one else is investing, the agency has had success with workshops that bring together people from diverse research backgrounds to create new areas of research. One recent example is bringing together researchers who work on two very different types of solar power, one that uses mirrors to concentrate sunlight and generate steam, and another that uses photovoltaics. Finding ways to combine the two could lead to solar panels that are both inexpensive and able to deliver power even when it’s not sunny—thus addressing the two biggest challenges of solar power.

Indeed, recent ARPA-E programs seem to be taking bigger risks than some of the ones announced a few years ago—one is even funding a project that will make batteries structural components of electric cars (see “Building Cars Out of Batteries Isn’t as Crazy as It Sounds”).

Although there are some short-term successes, like the Toyota retrofit, at the summit, there will be no shortage of longer-term projects that reflect ARPA-E’s ambition to truly transform energy technology.