Share article
Share article
Enjoy articles without ads?
Register for free and get unlimited access to all articles.
From bitcoin treasury flex to cashflow reality
The on-chain tell: listed miners are shrinking BTC holdings
Public miner Bitcoin holdings have fallen materially from peak levels. The latest figures point to a 15,096 Bitcoin decline from peak treasuries, and the bulk of that reduction is attributed to a handful of names: Core Scientific, Bitdeer, Riot Platforms, and Bitfarms.
That concentration matters. When selling is clustered among large operators, it tends to show up as repeated, size-able distributions rather than a smooth trickle. You don't need to romanticise miners as diamond-handed "apes" (retail traders who buy aggressively) to understand the impact. Miners sit at the source of new supply, and when they switch from holding to monetising, the market loses a natural absorber.
Why AI data centres are pulling capital from BTC
AI compute is the shiny new revenue line, but the pitch is not just vibes. There are structural reasons miners like it:
1) Better narrative for public markets
"AI infrastructure" gets a warmer reception from equity investors than "we run ASICs and hope number go up." Even if the economics are still evolving, AI spend looks like capex with a growth story, not just exposure to Bitcoin price.
2) Existing advantages, power and sites
Miners already control power access, land, cooling, and facilities. Retrofitting or repurposing sites for HPC hosting is often faster than starting from scratch, especially in regions where grid interconnects are a nightmare.
3) Diversification away from difficulty and fee uncertainty
The catch is that HPC buildouts are capital intensive. Racks, networking, cooling, contracts, and uptime requirements quickly turn into a funding question. When capital markets are picky, Bitcoin on the balance sheet becomes a liquid source of financing. [3]
What this means for BTC supply, it's the marginal seller that matters
Miners are not the only sellers in the ecosystem, and miner treasuries are not the entire market. Bitcoin's market cap at these prices sits roughly in the $1.3 trillion zone, so even tens of thousands of Bitcoin are not "infinite supply."
- Rally selling into strength, especially around liquidity clusters and prior highs.
- Less organic supply absorption, because the natural "mine and hold" bid disappears.
- More financing-driven flows, where Bitcoin is sold to fund capex schedules rather than because miners think the top is in.
This is where the scepticism kicks in. If AI buildouts are the new strategy, the market should assume more Bitcoin monetisation, not less, until those projects generate meaningful cashflow.
What to watch next (without getting lost in miner PR)
If you want to track whether this pivot actually pressures Bitcoin, ignore the buzzwords and watch the plumbing:
Miner-to-exchange behaviour
Sustained increases in miner outflows to exchanges are the classic pre-sell footprint. A few chunky transfers do not make a trend, but repeated distributions around the same cadence often do.
Treasury updates and financing notes
Public miners will often disclose monthly production and sales, plus whether Bitcoin is pledged. The key question is not "did they sell," it is "are they net sellers even after production?"
AI capex timelines
If miners are selling Bitcoin to fund buildouts, sales should correlate with construction phases and procurement cycles. Delays can cut both ways: less immediate selling, but also less revenue to replace it.
BTC price regime
With Bitcoin still well off its highs, balance sheet discipline matters more. If Bitcoin chops sideways, miners are more likely to fund operations and capex via sales. If Bitcoin trends strongly higher, they can refinance or raise equity on better terms and slow the selling.
The messy bit: AI hype does not guarantee miner upside
Plenty of miners are trying to pivot at the same time, and crowded trades get dodgy fast. AI infrastructure is operationally demanding, contracts can be lumpy, and the competitive set includes specialised data centre operators who have been doing this for years.
Risk box: what would invalidate the "more miner selling" thesis?
- Bitcoin breaks back toward all-time highs with strong spot demand, letting miners raise capital via equity or cheaper debt instead of selling coins.
- AI hosting generates real, recurring cashflow, reducing the need to monetise treasuries.
- Network economics improve sharply, for example, a sustained fee regime that boosts miner revenue enough to rebuild holdings.
Until one of those shows up in the data, assume the sector's centre of gravity is shifting: from "hold Bitcoin at all costs" to "sell Bitcoin to buy optionality in AI." That might be sensible corporate strategy, but for Bitcoin's market structure it sets up a simple outcome: more supply looking for bids.



