Why It Matters
Three of the largest publicly traded Bitcoin miners just told investors, in effect, that AI compute is worth more to them right now than mining bitcoin. Over the past week, TeraWulf, Hut 8, and IREN collectively signed close to $38.5 billion in long-term AI data center leases — more than the combined market capitalization of all three companies a year ago.
The Deals
TeraWulf signed a 20-year lease with Anthropic for a 401-megawatt campus in Hawesville, Kentucky, expected to generate roughly $19 billion in contracted revenue over its term. Initial capacity is due online in the second half of 2027, with the site reaching full build-out by early 2028.
Hut 8 signed a 15-year, $9.8 billion lease for its Beacon Point campus in Nueces County, Texas — a 352-megawatt facility built for AI training workloads rather than mining rigs.
IREN signed a five-year, $9.7 billion AI cloud contract with Microsoft, covering Nvidia GB300 GPU deployments across 200 megawatts of liquid-cooled capacity.
All three stocks rallied on the news: TeraWulf gained as much as 12.8%, IREN climbed more than 8%, and Hut 8 jumped nearly 10% — moves that had nothing to do with bitcoin's price that day.
Miner Valuations Are Decoupling
Analysts now say AI leasing headlines move these stocks more than bitcoin's price ever did. Listed miners could derive as much as 70% of their revenue from AI infrastructure by the end of 2026, up from roughly 30% earlier this year. TeraWulf, Hut 8, and IREN are no longer pure-play bitcoin miners in any meaningful sense — they are power and real estate companies that happen to still run mining rigs on part of their footprint.
This is not a surprise given the backdrop. Bitcoin Gate reported last week that roughly 20% of miners were operating at a loss, with breakeven costs near $78,000 against a bitcoin price closer to $64,000. AI leasing is the industry's answer to that squeeze: instead of running unprofitable hardware through a bear market, lease the power and the land to a hyperscaler for a guaranteed 15-to-20-year revenue stream.
These are not small side bets, either. A 15-to-20-year lease is a multi-decade balance sheet commitment, underwritten on the assumption that AI compute demand stays strong enough to justify the power buildout for two full decades. Miners are effectively trading one form of cyclicality — bitcoin's boom-and-bust price swings — for another: the AI infrastructure cycle, which has its own history of overbuilding ahead of demand. TeraWulf said its lease is expected to be supported by investment-grade credit, a signal it is trying to de-risk the commitment rather than lever up blindly, but the underlying bet is still a long one.
What It Means for Hash Rate and Security
There is a genuine argument this is good for the network. Diversified, contracted cash flow insulates miners from bitcoin's price swings and the shrinking block subsidy that follows every halving. A miner with a stable AI lease can keep facilities running and reinvest in efficient hardware instead of shutting down rigs during a downturn — which, all else equal, supports a more stable hash rate over time.
The counterargument is just as real. Bitcoin's security budget has always rested on a simple link: price goes up, mining becomes more profitable, more hash rate joins the network, security improves. As miners' revenue detaches from bitcoin's price and attaches instead to AI lease economics, that link weakens. Hash rate deployment decisions increasingly respond to data center demand and power contracts, not to bitcoin's price signal. Whether that trade-off is net positive or negative for long-term network security will not be clear for years.
For Bitcoin Investors, Not Miner Investors
The more immediate takeaway is simpler. Mining stocks have long been sold as a leveraged proxy for bitcoin's price — own the miner, get amplified upside when bitcoin rallies. That thesis is breaking down in real time. A TeraWulf or IREN shareholder is now, in large part, betting on AI data center execution, Nvidia GPU cycles, and hyperscaler demand — not on bitcoin.
If the goal is exposure to bitcoin's price appreciation specifically, buying and holding bitcoin directly remains the more predictable route. Bitcoin Gate's DCA calculator models what a steady accumulation plan looks like against 14 years of real price history, without the added variable of a miner's power contracts and GPU lease terms.
Bitcoin Gate Take
This is probably good news for miners' balance sheets and, cautiously, for the network's long-term resilience — diversified revenue beats a business model that goes underwater every time bitcoin drops 30%. It is bad news for anyone who bought mining stocks as a simple bitcoin bet. Those two things have been drifting apart for over a year, and this week made the gap impossible to ignore.