While Bitcoin fell roughly 17% through the first months of 2026, a basket of Bitcoin mining stocks rose more than 50%, with the best performers up over 70%. That divergence is not an anomaly. It is the clearest signal of the most important industrial transformation in crypto: Bitcoin miners are abandoning Bitcoin, or at least demoting it, to become artificial intelligence data centers.
The numbers are staggering. More than $70 billion in cumulative AI and high-performance computing contracts have now been announced across the public mining sector. Hut 8 signed a 15-year, $9.8 billion lease for a 352-megawatt Texas facility built to NVIDIA’s reference architecture. TeraWulf has locked in $12.8 billion in contracted AI revenue. IREN secured a $9.7 billion deal with Microsoft for 76,000 NVIDIA GPUs.
Industry projections suggest listed miners could derive as much as 70% of their revenue from AI by the end of 2026, up from roughly 30% today. The companies built to mine Bitcoin are becoming something else entirely, and they are selling their Bitcoin to pay for the transition.
The divergence that tells the story
The single fact that captures the whole transformation is the gap between miner stocks and the asset they were built to produce. In 2026, as Bitcoin slid on rising Treasury yields and hawkish Federal Reserve expectations, the companies that mine it went the other way. A tracked basket of crypto mining equities rose 56% year-to-date while Bitcoin itself fell about 17%, according to 10X Research. The individual leaders did far better. TeraWulf gained more than 73%. A handful of mining and AI-infrastructure stocks led the gains in the very weeks Bitcoin was bleeding.
The reason is straightforward once you see it. The market has stopped valuing these companies on how much Bitcoin they mine and started valuing them on how much AI computing capacity they can deliver. A miner that has signed multi-billion-dollar, 15-year leases with AI counterparties has a predictable, contracted revenue stream that looks nothing like the volatile, halving-exposed economics of Bitcoin mining.
Why mining stopped being good enough
Bitcoin mining was always a brutal business, and a confluence of forces in 2025 and 2026 made the AI alternative too attractive to ignore. Mining economics are punishing by design. Roughly every four years, the Bitcoin halving cuts the block reward in half, slashing miners’ primary revenue overnight unless the price rises enough to compensate. Miners compete in a zero-sum race for the same fixed pool of block rewards.
Then artificial intelligence created an almost perfectly matched opportunity. The AI boom produced explosive demand for data center capacity, and specifically for the two things Bitcoin miners already had in abundance: large-scale access to cheap power and the physical infrastructure to house and cool enormous racks of energy-hungry machines.
Instead of mining a volatile asset in a zero-sum halving race, a miner can sign a 15-year lease with a creditworthy AI counterparty for hundreds of megawatts of capacity, generating stable, contracted, dollar-denominated revenue with hosting margins that can exceed 25%.
Who is winning the pivot
Hut 8 has undertaken one of the most aggressive transformations in the sector. It signed a 15-year, $9.8 billion lease for its Beacon Point campus in Nueces County, Texas, a 352-megawatt facility designed to NVIDIA’s DSX reference architecture. The company’s posture says everything: Bitcoin is no longer a long-term strategic focus.
TeraWulf has been the credibility leader. It has signed HPC contracts totaling $12.8 billion, with deals anchored by Google-backed Fluidstack. IREN secured a $9.7 billion deal with Microsoft for 76,000 NVIDIA GB300 GPUs and holds zero Bitcoin in treasury.
One metaphor has spread across the sector: the “mullet data center.” Bitcoin mining runs in the back as a flexible, interruptible workload used to balance grid demand, while AI occupies the front where the multi-year contracts and stable margins live.
How they’re paying for it, and the risk that creates
The pivot is not free. The first source of funding is debt. IREN carries roughly $3.7 billion in convertible notes. TeraWulf has around $5.7 billion in total debt. If the AI demand softens or the buildouts run late, that debt becomes a serious problem.
The second source is more symbolic: the miners are selling their Bitcoin to fund the transition. Publicly listed miners have collectively reduced their Bitcoin treasuries by more than 15,000 BTC from peak levels. Core Scientific sold $175 million worth of Bitcoin in March 2026 to fund operational transitions.
There is also a concentration-and-oversupply risk. Because so many miners are pursuing the same pivot at once, there is a real possibility of overbuilding AI data center capacity relative to demand, which could compress the margins that make the strategy attractive.
What it means for Bitcoin
The most direct effect is on Bitcoin’s hashrate and network security. As miners divert power capacity from Bitcoin mining to AI workloads, computing power that would have secured the Bitcoin network goes to training and running AI models instead. Bitcoin recorded its first first-quarter hashrate drop in six years partly because of this diversion.
The second effect is selling pressure. The 15,000-plus Bitcoins that miners have sold to fund their AI transitions are real supply hitting the market. In a weak market, that miner selling is one more source of pressure on the price.
The deeper question is whether the pivot is reversible. The 15-year lease structures that dominate the new AI contracts make reverse migration economically irrational. The converted capacity is not coming back.
This article is for informational purposes and does not constitute financial or investment advice. Cryptocurrency markets are highly volatile. The figures and analysis described reflect data available as of June 5, 2026. Always do your own research and consult with qualified financial professionals before making investment decisions.
