A one-megawatt battery costs several million dollars, Brian Snyder told the room at Mining Disrupt Miami 2024, held June 24–26, 2024. A Bitcoin mine that switches off on request does much of the same job for the grid, and the grid does not pay for it. That was the case the Kaboomracks CEO made in a short keynote on power demand, renewables and where miners fit between them.
Snyder framed the talk for an audience that might not know the power business well. He covered what Bitcoin miners face on the grid today, where AI inference is adding demand, how wind and solar are changing the way the grid behaves, and why he thinks a flexible mining load can solve some of the problems that follow.
Demand Is Rising From Two Directions
He opened with a forecast from Dominion Energy for data center growth on its system. On a grid of 29.5 gigawatts, he said, Dominion expects data centers to reach 10 to 16 percent of total capacity, against roughly 1,200 gigawatts across U.S. markets. He added that the data is hard to pin down, with open questions about chip production and whether the build-out can keep pace.
Bitcoin's own energy use depends on price, and he walked through three scenarios to 2040. In a bearish case of $100,000 per coin, which he said probably won't happen, power use by mining would likely fall. A neutral case of $500,000 meant decent growth. His favored case, $2 million or more by 2040, meant a large expansion in energy demand. His question for the room was where that energy would come from if generation is not growing to match. Off-grid projects, such as flare gas mitigation that cuts methane emissions, were one answer he pointed to.
On AI, he split demand into training, which makes a model smarter, and inference, which he described as the energy used each time people interact with a model, such as a voice assistant at home. The chart he showed showed training falling and inference rising as it ran toward 2036. He noted that miners such as Core Scientific were already moving into AI, and said the result is more demand for cheap power.
The Baseload Problem
Snyder's main concern was supply. Nuclear and coal plants are shutting down, he said, and they are not being replaced with other baseload sources. Wind and solar are cheap once built, but variable. Turbines freeze in the cold, and solar produces only while the sun is out. Overproduction is common, and it is not easy to switch off. Transmission is the third limit: new high-voltage lines and transformers take years and a lot of capital, and he did not expect that to change without a breakthrough in superconducting materials.
He used the Texas winter storm as a warning. Texas, he said, had a relatively mild storm, and there were still fatalities. The Northeast came through the storms of the 1970s with nuclear and coal on the grid, though there were deaths and outages then too. If the Texas event happened there today, with temperatures of negative 20 to negative 40 degrees, he said the toll could run to thousands, if not tens of thousands.
Mining as an Elastic Load
His answer was to use Bitcoin mining as a load center that can be turned off. Grid-scale storage, in his view, is too capital-intensive to be practical today. Mines are built with private capital rather than government subsidies, and they can soak up surplus renewable power and drop off when that power is needed elsewhere.
it's also an elastic load, which means we can shut off the power when we need to return power to the grid so that people in their homes don't freeze to death at night.
Brian Snyder · 6:08
He described the trouble with baseload plants from the generator's side. A clean coal plant might average three cents per kilowatt hour in fuel and generation costs, which sounds good to a miner. But generators are paid on hourly or quarterly pricing, and on PJM there are periods when the price turns negative. A coal plant cannot ramp its generators up and down quickly; it can take a day or two. A mine can stop at once.
So not even zero cents, not even one cent, not even two cents, I actually have a negative power rate.
Brian Snyder · 9:21
That is where he sees the value. Miners can curtail when the grid needs power back, which turns renewable overproduction from a problem into a resource, as long as utilities set up programs that make the arrangement pay for both sides.
And so this is going to lead, it's basically, it's a free battery for the grid that the grid doesn't have to pay for.
Brian Snyder · 10:16
A Biomass Plant With a Mine Attached
Snyder then described a project he worked on in 2018, which he said never went ahead. A hundred-megawatt coal plant was being converted to biomass using the same boilers, footprint and infrastructure. Rather than pull all of the plant's power into a mine, the design kept 60 megawatts as baseload for the local community, with a capacity payment from the utility for providing it.
But you had a 40 megawatt Bitcoin load attached.
Brian Snyder · 11:19
By adjusting the fuel source, he said, the 40-megawatt mine could run 24 by 7, while the plant's remaining capacity was set aside for demand response: dropping load quickly to help stabilize the grid, as many operators already do on ERCOT. The projected payback came in far ahead of the original plan. Instead of an IRR measured over 15 to 20 years, he said, the project was looking at a matter of years, and with Bitcoin's later price moves probably about 18 months. The original projection was a 40% increase in plant revenue; with proper treasury management, he estimated it could have reached 400%. He also pointed to steadier operating costs, since a plant that shuts down when prices are low still pays its staff.
Power Access as an Asset
He closed on contracts and infrastructure. Many power purchase agreements are old, he said; ERCOT is further ahead than most, but agreements in general have not kept up with what the grid now needs from demand response. He called on utilities to adopt rate schedules that reward miners as a captive, flexible consumer.
For miners who own infrastructure, he said the power relationship itself will become valuable as AI and data center demand grows. That is why Kaboomracks, he said, is building its own facilities for customers to plug into. He said he would rather have the utility relationship secured in advance, since it adds to a company's enterprise value and could lead to a good payday.
that consumption of power, that access to the power is going to be worth a lot of money.
Brian Snyder · 13:36
Snyder ended by pointing attendees to the Kaboomracks booth, where his co-founders and business partners were on hand.
When Snyder spoke, Bitcoin traded at $61,272. On October 27, 2026, it was $79,612.