Summary:
Blockstream and Block (formerly Square) and Tesla have partnered to build a Bitcoin mining facility powered solely by solar plus batteries.
Bitcoin miners compete based on their energy costs, and new mining projects target energy costs of less than $0.03/kWh.
Based on 12 MWh of batteries, 3.8 MW of solar and 30 PH/s of ASICs, the Blockstream project will have energy costs of at least $0.15/kWh.
Even cutting battery costs in five, energy would cost about $0.1/kWh
Why thermal cycling and intermittency pose issues for Bitcoin mining with solar and wind:
I am a Bitcoin hodler since 2017. As I have previously written, I think the Bitcoin community is unrealistic about the potential for solar or wind power to be economic for mining Bitcoin, for two reasons:
The computers (ASICs) used for mining don’t like thermal cycling. This forces high levels of energy storage in order to provide a near continuous power supply.
Even without thermal cycling (in which case energy storage could be avoided), the high intermittency of solar and wind results in high downtime for the ASICs, which significantly reduces the payback period on capital deployed.
A financial model of the Blockstream project in Texas:
I have built out a financial model of the Blockstream / Block / Tesla project based on publicly available data. You can view the model (and copy it to play around with it) here.
Key takeaways:
The project appears sized to avoid thermal cycling of the ASICs. Based on information on the solar panels and the batteries, it seems it will be possible to run the ASICs with about 70%+ uptime. This is a lot higher than the 20-30% uptime there would be without energy storage. This makes sense given ASICs are known not to like thermal cycling.
Blockstream project energy costs are expensive. Bitcoin mining is all about having cheap energy. Miners with the cheapest energy can stay turned on for the longest when revenue falls (either due to falling Bitcoin price or rising network hashrate). From conversations with the top Bitcoin mining pools, I learned that new projects seek out energy cheaper than $0.03/kWh. By contrast, the – optimistically calculated – cost of power from this Blockstream project is $0.15/kWh.
Finishing remarks
I think it is fantastic that Blockstream/Block/Tesla have committed to doing this project and making the results public.
To become competitive, battery and solar costs would need to fall dramatically. Even if battery costs fell by 5X, the cost of energy would still be $0.1/kWh, which is not competitive.
I welcome improvements or corrections to my model, and look forward to seeing Blockstream’s published results.