Wind and solar at this point are cheaper than any other form of electricity, and are getting steadily cheaper still. Coal plants are getting shut down because it’s just not worth it to run them, and even natural gas is becoming noncompetitive. This is all to the good for air quality and slowing climate change, but it still leaves the problem of how to handle their variability. The Stanford Jacobson study says that we can meet 100% of our energy needs without fossil fuels or nuclear power, but it’s going to be a challenge.

But once again the engineers are coming through. There has recently been some good news on this front in three areas: the battery startup Ambri, the geothermal startup Eavor, and the research by the Australian National University on closed-loop pumped hydro-storage. Ambri has a scheme for dead cheap batteries with unlimited numbers of cycles, Eavor thinks they can build geothermal power plants almost anywhere, and ANU thinks that systems that pump water between high and low reservoirs to store energy can be built in tens of thousands of places. Let me talk about Ambri in this post and leave Eavor and ANU for subsequent ones.

Ambri was founded in 2011 by David Bradwell and Prof Donald Sadoway of MIT. I wrote about them here, Maniacal Energy Storage Schemes, and here, MIT on Climate Change. Sadoway thought that complex lithium-ion batteries would never be cheap enough to handle the terawatt-hours of storage needed. “If you want something to be as cheap as dirt, make it out of dirt,” he said. He taught the Intro to Solid State Chemistry course at MIT, which was recorded on EdX. One of the auditors happened to be Bill Gates, who arranged to meet him the next time he was in Cambridge. They talked about distance learning, and then Gates asked what else he was working on. Sadoway sketched out a scheme for using the different electro-potentials of liquid metals to make a battery, and Gates became their Round A investor. Teaching undergrads turns out to be good training for pitching VCs!

But they’ve had a long, hard road. This is what the hypesters of entrepreneurship rarely mention – everything will take longer and cost more than you expect. Their original chemistry didn’t work out, and the seals on the batteries failed at high temperatures. They laid off a lot of the staff in 2016, and started over. They’re now building cells using a calcium alloy anode, a calcium-chloride electrolyte, and granules of antimony:

The materials are poured into stainless steel boxes, put onto racks, and shipped as 1 MWh packs in insulated shipping containers:

They’re shipped cold, but heat themselves up to 500C when in operation in order to melt their materials. They’ve got an 80% round-trip efficiency, can start up in about a second, run at 500 to 1500V, and each container can output 250 kW. Li-ion batteries can catch fire and degrade with use, but these don’t. They’re claimed to be significantly cheaper than the $100 / kWh goal for Li-ion batteries, but of course don’t say how much.

The big recent news is that they’ve signed a contract for a 250 MWh system for a data center in Nevada: Ambri Inks Agreement With TerraScale’s Energos Reno Project To Deploy Proprietary Liquid Metal Battery Technology (Nov 24, 2020). At, say, $50/kWh that would be $12.5M. That’s not a heck of a lot, but gets them into mass production. The center will also have a big solar panel array, and is aimed at government and corporate customers who want secure, green computing. Massachusetts already has a center like that, the Mass Green High Performance Computing Center, which is powered by a dam in Holyoke MA, and has a 100 Gb/s fiber link to its founders: MIT, Harvard, BU, Northeastern, and U Mass. Siting this project in Nevada will give everyone on the West Coast similar low-latency access to green computing.

Still, everyone else in the world is working on Li-ion, and are already shipping tens of GWh of those cells. A proprietary solution may be better on the face of it, but has to compete with a thousand times as much ingenuity being applied to the mainline approach. But if Ambri doesn’t win, it’ll be because Li-ion got way better, and the world as a whole will be a winner. The prime value of capitalism is driving these kind of improvements through competition. As a citizen of Massachusetts, I’m hoping that Mass will be the place that delivers this breakthrough, but it’s going to happen somewhere.