Bitcoin Miners Sold Half Their BTC for AI. Did It Pay Off?
Publicly listed Bitcoin miners, including MARA Holdings, have sold nearly half their BTC since January, according to Glassnode. They are using proceeds to fund AI data centers. MARA sold 13,597 BTC for $1.1B. Four of six miners' stocks fell more than Bitcoin this year, with median loss of 41%. Bitcoin price is down 28% in 2023, trading at $63,947.
How this was made

The 30-second read
Why it matters
It links the behavior shift to added BTC spot supply during weak sentiment (Fear and Greed at 11) and then compares miner stock performance versus BTC over the same period.
Market read
Traders can use the report to reassess miner equity risk versus BTC, especially if continued BTC liquidation coincides with weak demand.
What to watch
The article relies on Glassnode-tracked reported holdings and compares YTD equity performance to BTC; it does not quantify each miner’s hedging, financing structure, or operational cost changes that could drive stock divergence.
Background
The piece uses Glassnode tracking of six listed miners’ reported BTC holdings to argue miners are transitioning from balance-sheet accumulation to selling to fund AI data centers.
Ticker impact
MARA sold 13,597 BTC for about $1.1 billion over six weeks to fund AI infrastructure, shifting from hoarding to selling.
Pressure on MARA shares as the market prices ongoing BTC liquidation for AI capex, especially if BTC demand stays weak.
The text ties MARA’s behavior change to continued BTC selling for AI build-out and links miner sales to added supply during “extreme fear.”
Riot Platforms is one of the few miners up on the year, with the article attributing gains to pushing into AI data centers.
Shares may remain supported relative to other miners if the market believes RIOT’s AI pivot reduces reliance on BTC sales.
The article provides year-to-date performance and a general AI-data-center rationale, but no specific RIOT BTC sale amount or filing detail.
CleanSpark is cited as one of the sharp underperformers, falling 47% this year alongside a shrinking BTC stack.
Downward bias if investors expect continued BTC supply from miners and CLSK’s equity beta to BTC remains high.
The article gives performance and general stack shrinkage but no CLSK-specific sale figure or new disclosure beyond the Glassnode tracking.
Hut 8 is reported up 5% on the year, with the article linking strength to AI data center expansion.
Potential relative strength versus other miners if the market continues rewarding AI pivot narratives.
The article lacks a HUT-specific BTC sale disclosure, so the catalyst is more narrative than incremental.
Cango is cited as the worst performer, down 89% this year, alongside a large reduction in reported BTC holdings.
Likely continued pressure if BTC liquidation persists and demand remains weak.
The article provides performance and general behavior change but no Cango-specific sale disclosure or new corporate action.
Market effects
Reinforces a sector-wide shift from BTC hoarding to funding AI capex, which can change how miners trade versus BTC.
Primarily impacts US-listed crypto-miner equities; sentiment spillover to broader crypto risk assets via BTC supply narrative.
Supports a global read-through that miner balance-sheet liquidation can weigh on BTC when spot demand is weak.
Counterpoint
Miner BTC sales may be partially offset by longer-term AI infrastructure cash flows, so equity underperformance could be more about timing and leverage than the AI pivot itself.
Key entities
- public_companyMARA Holdings
Largest BTC holder in the Glassnode chart; sold 13,597 BTC for about $1.1 billion over six weeks to fund AI infrastructure.
- public_companyRiot Platforms
One of the few miners up on the year, with the article attributing strength to AI data center expansion.
- public_companyCleanSpark
Reported as sharply down on the year (47%) alongside shrinking BTC holdings.
- public_companyHut 8
Reported up on the year (5%) with AI data center expansion cited.
- public_companyAmerican Bitcoin
Reported as down 84% on the year in the article’s comparison.




