Mining powered by "renewable" or "green" energy sources is a hot topic. For this episode, Compass talks with two engineers about the dynamics of the solar power industry and how can power bitcoin mining.
This discussion covers the incentives driving solar power's growth in the US, uptime and profitability considerations for miners, and how much hashrate will be powered by solar in the future.
Other ways to watch and read about this conversation:
Watch previous livestreams here.
Is solar power cheap? (timestamp)
- Yes, it’s cheap relative to other sources.
- Solar is also simpler because you don’t need to worry about fuel sensitivities.
- There are lots of ways to squeeze energy out of solar farms that aren’t happening now (e.g., curtailed energy, clipped energy)
Will miners choose solar because it’s cheap or because of ESG pressure? (timestamp)
- Some will be pushed because of the narrative.
- But ESG pressure isn’t necessary for solar to succeed.
- Risks to non-renewable miners aren’t non-existent though (e.g., recent statements from EU).
- Solar can be super cheap though, and that’s a strong incentive.
- A good narrative is: bitcoin mining helps renewables be more reliable.
The risks and data associated with machine maintenance and intermittent mining (timestamp)
- There isn’t much data available or an incentive to provide it.
- Constantly powering on and off the machines can probably reduce an ASIC’s lifespan, but we don’t know how severe the reduction will be.
- Miners who have operated load-balancing operations have a better idea of the effect on ASIC lifespans, but they haven’t shared their data.
Managing ASIC uptime with renewable energy sources (timestamp)
- At-home miners have huge potential for profitable solar-powered mining. Large-scale mining can be more challenging.
- Large-scale mining will either be intermittently mining or blend with solar plus something else.
- Scaling back capex by using less efficient (and cheaper) hardware is key for intermittent mining (e.g., Antminer S9s).
- Maintaining uptime would require a blended power source that draws from the grid when solar isn’t available, which is still pretty cheap.
- S9s at current market conditions can mine three times ($90) the value of a typical solar power purchasing agreement ($30).
Renewable energy and immersion mining (timestamp)
- Immersion mining is a rapidly growing trend.
- Miners can overclock their machines much more safely when immersed.
- 50% more hashrate at the same efficiency with immersion is a gamechanger.
- Immersion carries a higher capex, but overclocking can mitigate that.
- These optimizations might be unnecessary right now though, since lots of renewable energy is essentially free, so there’s no need to squeeze pennies right now.
- “Boatloads of money”
- Immersion economics can vary significantly by geographic location (e.g., Texas vs. Canada).
Solar energy development internationally versus the US (timestamp)
- Not sure about international dynamics; the industry is very complex.
- Solar energy development is exploding in the US though.
- Tax incentives are very strong.
- “Gigawatts of solar energy coming in everywhere.”
Will incentives for solar and other renewable energy last? (timestamp)
- Investment tax credits are a key driver for solar development.
- Solar farms already have power purchasing agreements to payback the capex from development.
- Combination of state and federal tax credits combined can create negatively priced energy in some places.
- Negative pricing will probably go away eventually.
- There are lots of places where solar is probably the only power source that could power mining (e.g., Australia).
How big is the carbon footprint to manufacture and build a solar farm? (timestamp)
- It’s tough to measure and compare fairly with fossil fuels.
- Solar probably has a smaller carbon footprint than coal. Natural gas is better, but still higher than solar.
- But every type of power has tradeoffs. And who’s to say carbon emissions are even bad?
What percentage of Bitcoin’s hashrate will be powered by solar before the next halving? (timestamp)
- This prediction needs to incorporate lower uptime and be weighted for solar variability.
- A very wild guess from Daniel: 3-5%
- Blake agrees, 5%. But a couple decades from now it will be much higher.