[dadiomov]

I studied energy at Stanford and joined the team that spun TerraPower (www.terrapower.com) out of Intellectual Ventures with Bill Gates' backing. I spent a lot of time looking at these nuclear batteries and economically it's very hard to see them working out as anything close to competitive to natural gas, coal, or wind. The amount of security, containment, etc. that you need for a large plant vs. a small plant doesn't vary much, so the costs per kWh don't scale down as you move down in total plant size, they scale up - and the idea as the techcrunch article suggests that you could sell just a "thermal" battery and let the utility or whoever add the actual steam-to-electric conversion on their own is laughable. Utilities don't do anything on their own.

I do hope these guys figured out something novel about the economics -- and I certainly applaud YC for going into energy.

[pjscott]

They're aiming at niches where they won't be competing directly with more conventional power plants. Their web site claims [1] that it's intended "for remote and distributed generation where energy costs can exceed 30 cents/kWh, and power is needed 24/7".

[1] http://www.upowertech.com/p/technology.html

[jddw]

Precisely right. Traditional utilities want something with a long operating track record, something hard for new designs to do off the bat. But going where the reactor is 5x cheaper than the next best option changes that. Plus many of these places want combined heat and power, so the design needs to be flexible on the thermal end.

[dadiomov]

there's definitely an opportunity in providing thermal at factories. lots of remote sites in India and China have their own purpose-built coal plants that could potentially use something like that...I have a harder time seeing the financing line up for serving remote villages that can't afford any power today. In those cases finance will always flow to smaller increment technologies, even if they're higher $/kWh, like solar + battery or diesel

[beefman]

The amount of security etc. required per site is subject to change. But even a 1GW site can be more cheaply assembled from 300 units coming off an assembly line than from one unit built in-place. The economics of this were discovered in the 18th century.