The Case for the Pivot
Give the AI pivot its due. The operational logic is excellent.
Bitcoin miners own the three things hyperscalers need most: land with power interconnects, cooling infrastructure, and teams experienced in managing high-density compute at scale. The skills transfer. The sites transfer. The people transfer. Converting a Bitcoin mining facility into an AI/HPC hosting site is, in many cases, a straightforward capital expenditure decision with strong unit economics.
And the numbers look compelling — in USD. GPU hosting contracts with hyperscalers offer more predictable revenue per megawatt, longer contract terms, and margins that don't collapse every time the Bitcoin network adjusts difficulty upward. Several publicly traded miners have already signed multi-year deals worth billions in aggregate.
Among the six publicly listed miners with HPC contracts, the CoinShares 2026 Outlook projects Bitcoin mining revenue falling from approximately 85% to under 20% of total revenue by late 2026. Bitfarms shareholders voted in March on renaming the company to Keel Infrastructure. CEO Ben Gagnon stated plainly that they are no longer a Bitcoin company but an infrastructure-first developer for HPC/AI data centres.
This isn't a fringe trend. It's the dominant strategic direction for the publicly traded mining sector.
The Denominator Problem
Here's where the steel-man meets the critique.
Every comparison between AI hosting and Bitcoin mining — revenue per megawatt, operating margin, IRR, contract value — is denominated in a currency that loses purchasing power every year.
At the experienced inflation rate of the last five years — roughly 5% annually — a USD revenue stream worth 100 today is worth approximately 61 in ten years. A multi-year GPU hosting contract that looks superior to BTC mining revenue today may look very different when measured in a unit of account that doesn't degrade.
The AI pivot isn't a business mistake. It's a monetary one.
The miners pivoting to AI have solved the operational problem — they've found a use for their infrastructure that generates cash flow. But they still have USD debt, and they're now servicing it with USD AI revenues. They've swapped Bitcoin volatility for a different kind of cyclicality — one that depends on hyperscaler demand, contract renewal rates, GPU generation transitions, and the pace at which large technology companies decide to in-source compute rather than lease it.
The Austrian Critique in Plain English
Cheap capital during the low-rate era drove miners to build infrastructure that only made economic sense at specific hashprices. When hashprice collapsed, that infrastructure didn't become good infrastructure. It became stranded capacity looking for a new story.
AI is the new story. And it might be a great one for some operators. But the structural question remains: is this infrastructure being valued on its intrinsic productive merits, or is it being repriced against a narrative that commands a premium today?
Core Scientific emerged from Chapter 11 restructuring with a $500M Morgan Stanley loan at approximately 7.8% — in USD — and over $10B in AI contracts with CoreWeave, denominated in USD. The company sold its remaining Bitcoin treasury and reorganised its entire capital structure around fiat-denominated revenue streams. From a traditional finance perspective, this is a successful turnaround.
From a monetary perspective, the company traded a volatile BTC revenue stream for a depreciating USD revenue stream. The capital structure is still built on fiat rails. The risks have changed in character — from hashprice and difficulty to contract renewal and hyperscaler concentration — but the underlying vulnerability is the same: obligations and revenues are both denominated in a currency whose purchasing power the operator does not control.
Who Actually Survives
The operators who endure across cycles aren't the ones who find the best pivot. They're the ones who never needed to pivot because their cost structure was positioned so far inside the breakeven line that the cycle was almost irrelevant.
Stranded gas miners running at $0.02–$0.03/kWh occupy this position. At those energy costs, breakeven hashprice is low enough that the first death — the operational one — can't reach them. They don't need to diversify into AI because their mining economics work even at the bottom of the cycle. Their competitive advantage isn't strategic flexibility. It's structural resilience.
This doesn't mean the AI pivot is wrong for every operator. For miners sitting on $0.05–$0.07/kWh power with next-generation data centre infrastructure, pivoting to AI hosting may be the rational decision. The sites have value that Bitcoin mining at current hashprices cannot capture.
But it does mean the pivot should be recognised for what it is: an exit from Bitcoin mining, not a solution to Bitcoin mining's structural problems. The miners who "pivot" are leaving the industry. The miners who solve the currency mismatch are the ones building its future.
The Fork in the Road
The Bitcoin mining industry is splitting into two camps.
One camp is migrating to AI/HPC, measuring returns in USD, and building capital structures around fiat-denominated contracts. This camp will succeed or fail based on the same factors that determine outcomes in any infrastructure hosting business — contract quality, counterparty concentration, technology refresh cycles, and capital cost.
The other camp is staying on the Bitcoin standard. These operators are optimising for energy cost, building capital structures denominated in BTC, and positioning for the 2028 halving as a feature rather than a threat. Their success depends on access to frontier-cost energy, sound credit architecture, and the long-term thesis that Bitcoin's purchasing power trajectory more than compensates for the volatility.
Both paths are rational. But they lead to fundamentally different businesses with different risk profiles, different return characteristics, and different relationships to the Bitcoin network.
The question isn't which path is better. The question is which denominator you believe in.
Ben Vincenzi is the founder of BTSF, a Bitcoin-native credit institution focused on structured lending to Bitcoin miners. He is the author of Beyond Digital Gold: Bitcoin and the Architecture of a New Monetary System.