From Mining Stage to Energy Stage: How AI Is Eating Bitcoin's Hashrate, and Why Drivechains Are the Only Way to Bring It Back

May 17, 2026 · 18 min read

At Bitcoin 2026 in Las Vegas this past April, the world's largest Bitcoin conference quietly made a change that most attendees did not notice. The "Mining Stage" — a fixture of every prior conference dating back to 2019, the place where mining executives, ASIC manufacturers, and pool operators gave talks about hashrate, difficulty, and the economics of proof-of-work — was renamed. It is now called the Energy Stage.

The renaming was not arbitrary. The conference organizers were responding to a reality that has become impossible to ignore: in 2026, the companies that have historically called themselves Bitcoin miners are no longer primarily Bitcoin miners. They are AI data center operators, with hashrate as a declining secondary business. The Mining Stage stopped being about mining because the speakers on it had stopped being about mining.

This is the most underreported security crisis in Bitcoin's history. And it is the clearest possible illustration of why drivechains are not optional. They are the only architectural response that addresses the underlying economic problem driving the pivot.

This essay is about what happened to Bitcoin's miners in 2026, what it means for the network's long-term security, and why Ecash — by activating BIP-300 from genesis — is the only chain positioned to bring them back.

What the renaming actually signals

A conference stage is a marketing artifact. Its name reflects what the organizers expect attendees to want to hear about. For seven years, from 2019 through 2025, the Bitcoin Conference's Mining Stage was a hub of substantive industry discussion: how to compete on electricity costs, how to manage ASIC depreciation cycles, how to weather price volatility, how to engage with regulators, how to coordinate on protocol upgrades.

In 2026, the stage's program shifted. The talks were no longer about mining as a primary business. They were about energy infrastructure as a primary business, with mining as one of several possible monetization paths for that infrastructure. The headline speakers were no longer pure-play Bitcoin miners. They were the executives of TeraWulf, Core Scientific, IREN, Cipher Mining, and Applied Digital — companies that, in some cases, no longer derive the majority of their revenue from Bitcoin mining at all.

The renaming acknowledged what the industry had already done. By April 2026, listed Bitcoin mining companies were projected to derive as much as 70% of their revenues from AI and high-performance computing by year-end, up from roughly 30% at the start of 2026. The "Bitcoin mining industry" had become, in any honest accounting, an "energy infrastructure industry that includes some Bitcoin mining."

The change is structural, not cosmetic. And it has profound implications for Bitcoin's security model.

The economics that drove the pivot

Bitcoin mining has always been a margin business. Miners earn the block subsidy plus transaction fees; they pay for electricity, hardware, cooling, maintenance, and capital costs. The margin between revenue and expenses determines profitability, and the profitability determines how much hashrate is willing to participate in securing the network.

For most of Bitcoin's history, this calculation worked in mining's favor. ASIC efficiency improved faster than difficulty rose. Bitcoin's price appreciation more than offset the periodic halving events that cut the block subsidy. New entrants kept the industry growing, even as marginal miners were priced out by more efficient competitors.

Then, starting in 2023 and accelerating through 2024 and 2025, the AI revolution created a competing use for the exact infrastructure Bitcoin miners had spent a decade building. AI training and inference workloads require massive amounts of electricity, dense cooling, sophisticated power management, and skilled operators. Bitcoin miners had built all of this. They had also built it in geographies with cheap power, near long-distance transmission infrastructure, with permits and grid connections already in place. They had, in retrospect, built the wrong product for the right substrate.

The economic comparison became impossible to ignore. By late 2025, Bitcoin mining was generating somewhere in the range of $50,000-$150,000 per megawatt of capacity per year, depending on price and difficulty. AI compute, meanwhile, was generating $1-2 million per megawatt per year for the same infrastructure. The math was not subtle. A miner who could pivot to AI was looking at a 10-20x revenue increase for the same physical assets.

The pivot began with the most flexible operators. Core Scientific signed a 12-year, $3.5 billion contract with CoreWeave to host AI compute. TeraWulf shifted significant capacity to AI workloads. IREN restructured its business around AI hosting. Applied Digital expanded its AI infrastructure aggressively while gradually winding down its Bitcoin operations. Cipher Mining followed similar paths.

By early 2026, the pattern had become industry-wide. Mining companies were announcing AI infrastructure contracts at a rate that exceeded their announcements of new mining capacity. The most successful "Bitcoin miners" of the 2020-2024 era were, by 2026, primarily known to financial markets as AI infrastructure operators with legacy mining businesses.

Why this is a security problem

Bitcoin's security model depends on hashrate being economically motivated to participate in mining. Mining is a competitive market. The total hashrate at any given moment reflects the marginal economic calculation of every operator — if mining is profitable, more hashrate joins; if it becomes unprofitable, hashrate exits.

What the AI pivot represents is a structural reduction in the willingness of major infrastructure operators to allocate their capacity to mining. They have not exited the industry in the traditional sense — they still own their ASICs, they still operate some of them, they still report mining revenue in their quarterly earnings. But the marginal allocation decision has shifted. When TeraWulf has a megawatt of capacity available and can deploy it either to AI hosting (revenue: $1-2M/year) or Bitcoin mining (revenue: $50-150K/year), the megawatt goes to AI.

The effect on hashrate is subtle but compounding. Hashrate has not collapsed dramatically. It has stayed roughly stable through 2025-2026, supported by the most efficient mining operations and by smaller operators in regions with very cheap stranded power. But the growth in hashrate that historically tracked Bitcoin's price appreciation has stalled. The natural expansion that would accompany price gains is being captured by AI workloads instead.

This is the security budget problem manifesting in a new dimension. The fee revenue collapse that Sztorc has been writing about for years was always going to be a problem as the subsidy halves toward zero. The AI pivot accelerates the problem, because it gives miners an alternative use for their infrastructure that does not depend on Bitcoin's economics at all. Even if Bitcoin's price doubles, the AI alternative will still pay more — and miners will continue to allocate their marginal capacity toward AI.

The deepest security threat to Bitcoin in 2026 is not regulatory, not technical, not political. It is economic. The companies that secure the network have discovered they can earn more by not securing the network. The only response to this is to give them a reason to come back.

What the Energy Stage talks were actually about

Walk through the program of Bitcoin 2026's Energy Stage and the shift becomes obvious. Sessions covered:

These titles, on their face, sound like the Mining Stage talks of previous years. But the content was different. "Strategic Mining" was about how mining can serve as a flexible load that complements AI workloads — running when AI demand is low and pausing when AI demand is high. "Lord of the Rigs" was about how ASIC manufacturers are pivoting to GPU-adjacent designs that can serve dual purposes. "Game Theory of Miner Centralization" addressed how the consolidation of mining into a small number of AI-pivoted megacompanies changes Bitcoin's security model.

The implicit message of every talk was that mining is no longer the primary business. Mining is now a secondary business that complements the primary business of AI hosting. The Mining Stage became the Energy Stage because the people on it were no longer there to talk about mining as their main job.

A particularly telling moment came in a session featuring Fred Thiel, formerly CEO of MARA (the company once called Marathon Digital Holdings, the largest pure-play Bitcoin miner). Thiel, speaking to WIRED in 2025, had cast doubt on whether Bitcoin miners could pivot to AI as seamlessly as they wanted. By 2026, he was on the Energy Stage discussing exactly how that pivot was happening at scale. The transformation from skeptic to participant happened in about eighteen months, which is roughly how long it took the AI economics to overwhelm the mining economics.

Why drivechains are the only structural answer

The standard responses to the security budget problem are inadequate. "Price will save us" assumes Bitcoin's appreciation can outpace the AI competition for the same infrastructure, which the math does not support. "Fees will rise" assumes scarcity on the L1 will eventually drive fee revenue to levels that compete with AI hosting margins, which would require L1 fee rates so high that ordinary users would be priced off the chain entirely.

The only structural response that addresses the underlying economic problem is to give miners additional fee revenue streams that come from sources other than the L1. This is precisely what drivechains do.

A miner who secures both Bitcoin's L1 and an Ecash drivechain ecosystem earns fees from every sidechain that achieves any meaningful adoption. Thunder, the high-throughput payments sidechain, generates fees from every micropayment, remittance, and merchant transaction. ZSide, the privacy sidechain, generates fees from every shielded transaction. Hivemind, the prediction market sidechain, generates fees from every bet placed on every market. BitNames, the identity sidechain, generates fees from every name registration. CoinShift, BitAssets, Photon — each one contributes to the miner's revenue stack.

The math compounds quickly. A privacy sidechain processing a million transactions per day at fifty cents each generates $180 million per year in fees, paid entirely to miners via blind merged mining. A prediction market doing Polymarket-scale volume generates tens of millions more. A payments sidechain at Visa-scale throughput generates billions. Each sidechain that finds product-market fit adds another revenue stream, all flowing to the same hashrate that secures the underlying L1.

This is the architectural counter to AI. It is not a matter of competing with AI on margins — that competition is structural, and Bitcoin loses it. It is a matter of adding enough independent revenue streams that the combined margin from Bitcoin + drivechain fees becomes competitive with AI alone. The miners do not have to choose between Bitcoin and AI. They can do both, and the drivechain economics are what make the Bitcoin half worth doing at scale.

What Ecash specifically changes for miners

The argument for drivechains has existed since 2015. What is different in 2026 is that the political process inside Bitcoin has comprehensively demonstrated its inability to ship them. The soft fork that would have brought drivechain economics to Bitcoin miners has not happened, will not happen on the existing chain, and is therefore not available to the AI-pivoting miners as a reason to stay engaged with Bitcoin proof-of-work.

Ecash changes this. The August 2026 fork activates BIP-300 and BIP-301 from genesis. The drivechain economics become real on Ecash, on day one, regardless of what Bitcoin's political process eventually does.

This creates an immediate opportunity for the very miners who have been pivoting to AI. Ecash uses the same SHA-256d proof-of-work as Bitcoin. The hardware is fully compatible. The operational expertise is fully transferable. A miner who pivots a percentage of their capacity to Ecash mining can immediately access drivechain fees that do not exist on Bitcoin. The question for each operator becomes: do I leave that revenue on the table, or do I capture it?

The answer is almost certainly going to be the latter, for a simple reason: the marginal economic calculation that drove the AI pivot is the same calculation that will drive Ecash adoption. A miner allocating capacity between Bitcoin, Ecash, and AI hosting will look at the revenue per megawatt of each option and allocate accordingly. Bitcoin alone loses to AI. Bitcoin + drivechain fees on Ecash might compete with AI, depending on how rapidly the sidechain ecosystem grows. The miners who allocate early to Ecash will be the ones best positioned to benefit from sidechain fee growth.

The same economic logic that drove the AI pivot will drive Ecash adoption. Miners do not have to choose between Bitcoin's monetary properties and AI's superior economics — they can have both, on a chain that activates the architecture Bitcoin refused to activate.

What this means for the August 2026 fork

The fork is happening in a moment of genuine crisis for Bitcoin's security model. The Mining Stage has become the Energy Stage. The largest miners are no longer primarily miners. The fee revenue available on Bitcoin's L1 is structurally insufficient to compete with AI hosting margins. The political process that could have addressed this through drivechain activation has failed.

Ecash forks into this environment with a value proposition that should resonate strongly with miners specifically:

The combination is potent. A medium-sized AI data center operator can allocate, say, 10% of their capacity to Ecash mining at launch, capture disproportionate early block rewards from the difficulty reset, and accumulate Ecash positions that benefit from sidechain fee growth over the subsequent years. The cost of trying is minimal — existing ASICs, existing infrastructure, existing expertise. The potential upside is meaningful, especially if any of the drivechain sidechains (Thunder for payments, Hivemind for prediction markets, ZSide for privacy) achieves the kind of adoption their designs suggest.

For the operators who have been pivoting to AI, Ecash represents a hedge. If drivechain economics work as Sztorc has predicted, Ecash mining will become a meaningful revenue stream over the coming years. If they do not work, the cost of having participated is minimal. The asymmetric upside makes participation rational for any sophisticated mining operator.

The deeper structural shift

There is one more layer worth addressing. The Mining Stage to Energy Stage renaming reflects more than just an industry pivot. It reflects a shift in how Bitcoin's relationship to physical infrastructure is understood by the people who own that infrastructure.

For most of Bitcoin's history, mining infrastructure was understood as Bitcoin-specific. ASICs were single-purpose. Cooling systems were optimized for ASIC density. Power contracts were structured around mining's specific load characteristics. The infrastructure existed because Bitcoin existed.

In 2026, that relationship has inverted. The infrastructure now exists because energy markets exist, and mining is one of several monetization paths for it. The same megawatts can be deployed to mining, to AI training, to AI inference, to grid services, or to whatever new compute use case emerges next. The infrastructure operators are not committed to mining. They are committed to revenue optimization across whatever workloads pay best at any given moment.

This is, in some ways, a healthier industrial structure than the pure-play mining model that preceded it. It means that energy infrastructure built for Bitcoin mining can be repurposed when economics shift, rather than becoming stranded assets. It also means, for Bitcoin specifically, that hashrate is increasingly elastic in a way it was not before. The hashrate will be there if the economics support it. The hashrate will quietly disappear if the economics do not.

Drivechains are the mechanism for making sure the economics support it. By giving miners multiple parallel fee revenue streams from sidechain activity, drivechains keep the marginal economics of mining competitive with the AI alternative. Without drivechains, mining loses the marginal allocation decision more often than it wins, and hashrate gradually erodes. With drivechains, mining wins the marginal allocation decision often enough to keep hashrate growing alongside the broader compute economy.

What the Energy Stage talks should have been about

Imagine the Bitcoin 2026 conference if BIP-300 had activated in 2018, when Sztorc was first proposing it seriously. The Mining Stage talks would have been about how to capture drivechain fees from a thriving sidechain ecosystem. The discussion would have been about which sidechains were attracting the most users and therefore generating the most fees. The competition between mining operators would have been about who could best serve the sidechain ecosystem's needs.

Instead, the Energy Stage talks are about how to pivot away from mining toward AI. The competition between operators is about who can best serve OpenAI, Anthropic, and the other AI labs. The fee opportunity that drivechains would have provided is not on the table because the political process refused to put it there.

This is the cost of activation theater, made visible in the most concrete possible way. The mining industry that exists in 2026 is the mining industry Bitcoin's political process produced. It is structurally smaller than it would have been, structurally less committed to Bitcoin specifically, and structurally more vulnerable to being captured by adjacent industries.

Ecash is the chance to reverse some of this damage. Not all of it — AI is not going away, and the infrastructure operators who have built AI businesses will continue running them. But the marginal allocation decisions, going forward, can favor Ecash over pure Bitcoin if Ecash's drivechain economics work as designed. Each percentage of mining capacity that allocates to Ecash instead of pure AI is hashrate that secures the chain Sztorc has been arguing for since 2015.

The invitation to miners

If you are a mining operator who has been listening to the Energy Stage talks and quietly pivoting to AI, this essay is, in part, an invitation.

You do not have to choose between Bitcoin and AI. You can mine Ecash alongside your AI workloads, capture drivechain fees that do not exist on Bitcoin, and hedge your bets on which architecture wins the long-term competition for the cryptocurrency economy. The infrastructure investments you have made are not wasted by participating in Ecash mining — they are enhanced, because they generate revenue from multiple parallel streams instead of just one.

The Bitcoin community spent a decade refusing to ship the architecture that would have kept you primarily focused on mining. The result is the Energy Stage. The result is your pivot to AI. The result is a security situation for Bitcoin that nobody at the conference was willing to discuss honestly. Ecash is the architectural response that arrives, finally, at the moment when miners need it most.

The fork happens at block 964,000, approximately August 2026. The difficulty resets at launch. The drivechain infrastructure activates from genesis. Thunder, Hivemind, BitNames, ZSide, and the other planned sidechains begin generating fees within weeks of launch. The miners who allocate early will capture the disproportionate block rewards and the early-stage fee opportunities. The miners who wait will capture the steady-state economics after the market sorts out which sidechains have product-market fit.

Either approach is rational. What is not rational is continuing to pretend that Bitcoin mining alone can compete with AI hosting on the margins that exist in 2026. The math does not work. The Energy Stage exists because the math does not work. Ecash is the proposal that adds enough additional math to make it work again.

The Mining Stage was renamed because the speakers stopped being miners. The new chain — the one that activates the architecture that would have kept them miners — is finally launching. It is not too late to allocate some hashrate to it.

That is what Ecash means for the AI-pivoting miners of 2026. The Bitcoin you helped secure is becoming structurally less secure because the economics no longer favor mining it. The Ecash that activates next month is the chain where the economics will favor mining again, because the drivechain fees will be there. You built the energy infrastructure. The decision about what to run on it is yours. Drivechains give you a better reason to run mining workloads on at least some of it than Bitcoin alone has been able to provide for years.

The Energy Stage was right about the present. Ecash is the bet on a different future.