Are these projects high-risk enough?
Some of them will fail now. And we'll terminate them and put money where it either works or just put it back in the treasury. But some of the failures may happen beyond ARPA–e. It works now, but if you really try to scale the volume or the cost of fuel production, that's beyond ARPA–e. It may fail then and we may not know.
Even 1366 [Technologies, a solar photovoltaic company]—we may call it a success right now because they got private sector funding, they met their milestones, etcetera. But who knows whether that thing is going to be deployed everywhere in the world? That would be big success, producing tens of gigawatts of power using 1366 modules. That would be real success. Right now it's an early success. As I said in my talk these are early signs of success when the private sector comes in but that doesn't mean full success yet. From ARPA–e's perspective, at the end of the two years they will be a success, but there's a long way to go. So I would say we'll see how these red alerts go. There may be some terminations but there may be failures downstream [as well].
Is the ARPA–e timeframe too short to determine whether something is a success or failure?
That's possible. We tell our folks that, look, if something's not working and we terminate a project—don't take it personally. Come back again. We want to help innovators come back again with better ideas, and we will try to support you. Which is why we bring the finalists here [to Washington, D.C.,] and showcase them. These are people we could not fund. These are not failures.
Projects have two or three years with yearly go/no-go milestones. If some manufacturing process just does not work based off an air-conditioning cycle, then, gosh, there's no reason to pursue that until they come up with an idea that will overcome the barrier they are facing. When something's not working, they've faced a barrier that was unplanned. They were hoping to overcome another barrier, which they did, but they face another barrier and have no plans for it. Come back with another idea to overcome that barrier, and then we'll look at it again.
What about cost constraints?
In this case we are trying to create fuel. We know what the target price is. If you can get it at $2 per gallon you're in good shape…. The operational model is that of a Manhattan Project. You had Oppenheimer who was a deep scientist who [was] coordinating the different areas that need[ed] to be coordinated to get there. But there is a target out there in terms of cost [for ARPA–e]. Yes, we're not creating the bomb. We're taking the operational model, getting industry on board as a constraint on cost, and eliminating things that will not make the cut. If you have a catalyst based on ruthenium, that's not going to scale. You might as well get it out of iron then. Then focus on the iron…. The operational model is [also] like DARPA: We hire people, they're there for a limited amount of time, and then they have to leave. We create workshops. I've been funded by DARPA for more than a decade [in the past]. No one ever asked me how much it costs. No one—ever. You bet we ask that in ARPA–e, because it's a different ball game out here in the energy sector. Just because we take the operational model doesn't mean that we're going to create the bomb out there with no cost constraint. It's a different goal.
What's next for ARPA–e in terms of new programs, assuming you continue?
We have the world's largest reserves of natural gas so, yes, I want to go in that direction at some point. There's a lot of work to be done.