Bitcoin Miners’ AI Revenue Set to Match Mining by Year-End
For ETF investors, the key question is not whether bitcoin miners can diversify into AI, but how that shift changes the risk profile of the underlying equity basket. A miner with an AI contract still sits inside a business model that is highly sensitive to power access, financing costs and grid policy. That means the ETF exposure is evolving from a pure crypto proxy toward a mix of digital infrastructure, speculative growth and energy-intensive compute demand.
The composition angle matters. If more revenue comes from longer-dated AI and high-performance computing contracts, the companies may look less exposed to bitcoin’s day-to-day swings. But that can also increase concentration in a different way: a small number of tenants, sites and power agreements may begin to drive valuation. For an ETF, that makes index rules, rebalancing cadence and constituent screening more important than the headline theme.
Investors should also weigh structure and implementation details. The article highlights a strategy broadening toward neo-clouds, which can change sector overlap and correlation with other AI infrastructure names. In a fast-moving theme, that may improve diversification at the portfolio level, but it also raises the chance that the fund becomes less intuitive to own if the holdings drift away from the original mining thesis.
- Look at how the index defines eligible revenue sources.
- Check whether concentration limits reduce single-name or customer risk.
- Remember that valuation can re-rate quickly when power access becomes scarce.
- Use the ETF as a satellite exposure, not a stand-alone thesis.
Bitcoin miners are on track to earn about as much from artificial intelligence as from mining by year-end, according to two CoinShares portfolio managers.
Key Takeaways:
- Mining stocks fell more than 30% in October, as worries about a major AI customer weighed on the group.
- Demand to connect new projects to the U.S. grid is twice the size of its installed capacity.
- Several states have started tightening rules for large power users, with New York pausing permits for large loads.
Roxanna Islam, head of sector and industry research at VettaFi, hosted the CoinShares-sponsored webcast Mining’s new math: Power, compute, and sustainability on Wednesday. Her guests were CoinShares ETF portfolio manager Bill Cannon and equity index manager Alex Schmidt.
Miners reported $512 million in AI and high-performance computing revenue in the second quarter, or about 32% of the total, Schmidt said. Roughly 70 contracts ramping up later this year should add about $400 million in AI revenue per quarter.
See more: Bitcoin Miners Trade Rigs for AI as Power Becomes the Prize
The April 2024 halving cut miners’ bitcoin reward in half and squeezed margins. At the end of the second quarter, producing one bitcoin cost about $75,000 on average, Cannon said, citing CoinShares research. Bitcoin’s price, meanwhile, sat just under $60,000.
AI pays more for the same power. AI and high-performance computing work earns about $1.5 million per megawatt a year, versus roughly $500,000 for mining, Cannon said.
Steadier, long-term contracts have also lifted valuations, according to Schmidt. Miners’ enterprise value-to-sales ratio rose from four to five times, a year or two ago, to eight to 12 times.
Why Miners Hold the Grid Advantage
Power access is the main draw. Projects waiting to connect to the U.S. grid total about 2,600 gigawatts, double its 1,300 gigawatts of installed capacity, Cannon said.
States are also tightening rules for large power users. New York, for one, became the first state to pause permits for loads of 50 megawatts or more, Schmidt said.
“The scarce asset is no longer land or GPUs, but energized interconnections, and miners own exactly that,” Schmidt said.
Share prices have still been bumpy. Some miners fell more than 30% in October, while bitcoin held between $80,000 and $86,000, Islam noted.
Schmidt tied much of the drop to worries about CoreWeave, Inc. (CRWV), a key AI customer for some miners. Heavy selling pushed yields on some CoreWeave bonds above 15%.
Cannon manages the CoinShares Bitcoin Mining and Digital Power ETF (WGMI), which he said widened its strategy in August to include more neo-clouds, or firms renting out AI computing power.
“It’s an energy story, but with an AI revenue line and a crypto option,” Cannon said.
Converted sites may not return to mining soon. Schmidt said outfitting a site for AI tenants costs about $10 million to $15 million per megawatt, versus $1 million to $1.5 million for mining. That gap means bitcoin would have to be “much higher than it is now” to justify switching back.
Enjoyed this article? Sign up for our newsletter to receive regular insights and stay connected.