Waste or work - the energy case for bitcoin mining beyond price speculation

In late 2025, Elon Musk broke a long silence on bitcoin with a single line that travelled across timelines within hours: "energy is the true currency". Speaking on a podcast with the Indian investor Nikhil Kamath, and echoing a post he had made on X that October, he argued that governments can print fiat at will but no one can print energy, and that bitcoin is grounded in physical reality because mining converts real electricity into a fixed issuance schedule.

The quote is a useful hook. It is also the least interesting part of the story.

The monetary framing, bitcoin as monetised energy, has been familiar inside the industry for years, and the gold advocate Peter Schiff was quick to dismiss the asset as fundamentally hollow in the same news cycle. The shift worth the attention of anyone working in energy or capital markets is not that argument at all. It is operational, and it has almost nothing to do with the price on a screen.

Power markets are quietly starting to treat bitcoin mining not as a wasteful curiosity but as a controllable industrial load, a buyer for energy that would otherwise be discarded, and in a growing number of cases a source of usable heat. This article looks at the evidence behind that shift, the numbers that support it, and the one development in 2026 that may be undoing it.

TL;DR

• The hook. Musk reframed bitcoin as energy-backed money in late 2025. The more useful story is what mining now does for the grid, not what it signals about price.

• The flip. The standard criticism, that mining wastes electricity, rests on interruptible demand, which is exactly the property modern grids now pay for. In Texas, miners curtail to near zero within seconds and are treated as controllable load.

• Waste as fuel. Curtailed wind and solar, and flared or vented methane, can run the same machines. Research suggests methane mitigation through mining could remove a meaningful share of global emissions, though critics dispute the framing.

• Even the heat. Roughly nine tenths of the electricity a miner draws leaves as heat. Pilots in Canada and the Nordics now sell that heat to greenhouses and district heating networks.

• The catch. The 2026 pivot to AI and high-performance computing rewards firm, always-on power. The flexibility that makes mining useful to a grid is the first thing AI hosting takes away.

• The question. Is bitcoin mining a cost the grid tolerates, or infrastructure the grid relies on. The honest answer in 2026 is that it is becoming both, in different places, at the same time.
  

What did Musk actually say, and why does it matter?

Musk's argument runs through physics rather than finance. Energy, he said, cannot be legislated into existence or printed by decree, and proof of work ties a digital asset to that constraint by forcing miners to spend measurable electricity to secure the network. He extended the point into speculation about a post-scarcity future in which money fades and energy generation becomes the effective measure of value, invoking the Kardashev scale and the science fiction of Iain Banks.

That is the monetary thesis. It is interesting and it is contested, and it does not change anything that a grid operator does on a hot afternoon. The development that does is the one the headline never mentioned. For years the standard objection to bitcoin was that mining burns electricity for no productive purpose. The property sitting underneath that objection, demand that can be switched off without warning, is the property that energy systems have started to value most.

Why is interruptible demand suddenly an asset?

A bitcoin miner can cut its power draw to near zero within seconds, with no damaged equipment and no lost product. Almost no other large industrial consumer can do that. In Texas, the grid operator ERCOT now classes mining sites as controllable load resources, eligible to provide ancillary services and to respond to grid stress in anything from five minutes to under a minute.

The scale is no longer marginal. According to Spark's 2026 review of mining's energy mix, crypto mining demand in Texas reached roughly 4,288 megawatts by November 2025, with projections beyond 5,300 megawatts by 2027, and ERCOT had approved around 9,500 megawatts of large flexible load capacity by the end of 2025. The reliability record is concrete rather than theoretical. Miners reduced
consumption during Winter Storm Uri in 2022, the 2023 summer heatwaves and Winter Storm Heather in 2024. The economics are concrete too: in August 2023 ERCOT paid a single miner 24.2 million dollars in curtailment credits, plus a further 7.4 million dollars through demand response programmes, for being willing to stand down when the grid was tight.

The wider context matters for how this is read. ERCOT's interconnection queue now holds more than 438,000 megawatts of proposed demand, around 90% of it from data centres, and Texas regulators have approved a new allocation framework to manage the crush. A load that can throttle itself on command is not a problem in that environment. It is part of the solution.

How does mining change the economics of renewables?

Wind and solar are intermittent, and when they overproduce, surplus power is curtailed and wasted because the grid has nowhere to put it. A buyer that runs on that surplus and switches off the moment power becomes scarce improves the economics of building more clean capacity in the first place. Mining can be that buyer, which is one reason operators cluster in west Texas, where wind and solar corridors overlap.

The energy mix has moved in step. The Cambridge Centre for Alternative Finance, in its April 2025 industry report, put the sustainable share of the sampled network at about 52%, and treated that as a floor rather than a definitive global figure, since the unsampled portion in Russia, China and Central Asia probably leans more fossil.

The climate analyst Daniel Batten, using a different dataset, puts the figure higher at around 57%. Either way the direction is consistent: coal, which supplied over a third of mining power in 2021, now supplies far less, and nuclear has roughly doubled its share to close to a tenth. The honest caveat is that the grid benefit is real only where miners actually sit on surplus or waste power and actually curtail. A miner drawing firm fossil power at full load through a heatwave delivers none of it.

Can mining actually cut emissions? The methane case

This is where the argument becomes counter-intuitive, and where it needs the most care. Methane is a short-term super-warmer, tens of times more potent than carbon dioxide over a couple of decades when it escapes unburnt, and the UN Environment Programme has described cutting it as the strongest lever available to slow warming over the next 25 years. A great deal of it leaks from oil sites as flared or vented gas that never fully combusts. Satellite work by the Environmental Defense Fund in the Permian Basin found official flare estimates understated by a factor of around 3.5.

Bitcoin mining is location-agnostic, mobile and interruptible, which is the unusual combination that lets it sit on a remote gas flare or a landfill and turn otherwise leaking methane into electricity. Routing that gas through a generator combusts it, cutting its warming impact by up to about 90% compared with venting.

Daniel Batten's modelling argues that, taken to its potential, methane mitigation through mining could eliminate up to 5.32% of all global emissions by 2045, and that a few dozen mid-sized venting landfills running mining loads could in principle make the entire bitcoin network carbon-negative. Operators such as Crusoe have been cited by the World Economic Forum for landfill and flare-gas work, and a peer-reviewed model of a 1.14 megawatt landfill project estimated mitigation of more than 2,000 tonnes of methane.

The counter-case deserves equal space, because a serious reader will already be sceptical. Groups including the Sierra Club call this greenwashing, arguing that paying to consume flare gas hands oil and gas producers a new revenue stream and prolongs the underlying activity. The mitigation figures are modelled maxima, not realised outcomes.

The sustainability percentages vary with methodology and sampling. None of that is settled, and the responsible framing is that mining has a credible and partly demonstrated role in methane mitigation, not that it has solved the problem.

What happens to the heat?

Almost all of the electricity a miner consumes leaves the machine as heat. Historically that heat was vented to the air, which is the literal definition of waste. A cluster of 2026 projects treats it as the product instead.

Hardware maker Canaan and Bitforest launched a 3 megawatt pilot in Manitoba in January 2026, running 360 liquid-cooled servers whose captured heat raises water above 75 degrees and preheats the boilers of a commercial tomato greenhouse. By May, Canaan had a Nordic district heating project targeting 8 megawatts and around 2,800 homes, with roughly 2 megawatts already live and supplying hot water to residents.

The miner MARA reported integrating mining into two Finnish district heating systems in under 30 days, displacing legacy fuels such as peat, oil and wood. A Latvian firm, Power Mining, builds shipping-container units it says can mine close to ten bitcoin a year while heating about 2,000 homes, with the first units bound for Scandinavia. The model borrows directly from conventional data centres in Finland and Sweden, which already warm residential districts through municipal heat grids.

Heat reuse does not erase mining's footprint. It improves how efficiently that energy is used, and it changes the optics. When a heating node can be serviced without shutting down the heat source, mining begins to behave like infrastructure rather than a side project bolted onto a warehouse.

The catch: is the AI pivot undoing the energy case?

Here is the part the bullish narratives tend to leave out, and it is the most important development in the sector right now.

The dominant story in mining through 2025 and 2026 is not flared gas or district heat. It is the pivot to artificial intelligence and high-performance computing. Public miners have announced more than 70 billion dollars in cumulative AI and HPC contracts. Core Scientific signed roughly 10 billion dollars of capacity through

CoreWeave and reported colocation revenue that already runs at more than double its mining revenue. IREN agreed a 9.7 billion dollar deal with Microsoft for 76,000 of Nvidia's GB300 GPUs. Riot debuted data centre revenue and is building a campus designed for up to a gigawatt of power. Industry projections suggest listed miners could draw up to 70% of their revenue from AI by the end of 2026, for the simple reason that AI hosting earns far more per megawatt than mining a block subsidy that halves every four years.

The problem is structural, not financial. AI training and inference require firm, consistent, around-the-clock power. They are close to the opposite of an interruptible load. As mining capacity is repurposed for AI, the very flexibility that makes a site valuable to a grid is the first thing to disappear, even as total consumption rises.

The two bullish arguments for mining therefore pull against each other. The energy-utility thesis is strongest for pure-play miners that stay interruptible and sit on waste power. The AI thesis rewards the operators that convert to firm load and stop curtailing. Capital is currently flowing toward the second, which means the demand-response and waste-gas case may narrow to the operators who deliberately stay out of the AI race.

So is bitcoin mining a cost or infrastructure?

The label depends on which operator, which grid and which year you are looking at, so it is worth holding two scenarios rather than picking a winner.

In the first, pure-play and hybrid miners deepen their role as flexible load, methane mitigators and heat suppliers. Mining reads increasingly as grid infrastructure and waste-to-energy, and the energy criticism that defined the last decade reverses into an energy argument. In the second, the AI pivot dominates the listed sector, fleets convert to firm power, flexibility fades, and the grid case contracts to a smaller cohort of operators who choose interruptibility on purpose. Both are happening at once.

What is no longer credible is the flat claim that mining is simply electricity burned for speculation. The power it draws is increasingly doing measurable work: holding grids together, monetising gas that would otherwise leak and heating homes and greenhouses. Musk's line about energy as the true currency was a philosophical flourish. The more grounded version is a question for the people who plan power systems and allocate capital. When an industrial load can stabilise a grid, consume waste and deliver heat, do you still file it under cost, or under infrastructure.

Key terms

• Controllable load resource (CLR). A consumer that can adjust its power draw quickly on instruction from the grid operator, and is paid to do so.

• Demand response. Programmes that pay large consumers to reduce consumption during periods of grid stress, improving reliability without building new supply.

• Curtailment. The deliberate reduction or shutdown of power, either of generation that cannot be used or of a flexible load when power is scarce.

• Stranded or flared gas. Natural gas, often methane, produced where there is no pipeline to carry it, typically burned off or vented at the site.

• Proof of work. The mechanism by which bitcoin miners expend electricity and computation to secure the network and earn newly issued coins.

• Large flexible load. A grid category, used by ERCOT, for very large consumers able to vary demand, including mining sites and data centres.
  

About Bitcoin Poland Conference

Bitcoin Poland Conference is Poland's first bitcoin-only conference, taking place at the Poznań Congress Center from 4 to 6 October 2026. Co-hosted with Invest Cuffs and Cashify, it brings together the people building bitcoin's monetary, financial and energy infrastructure for three days of analysis and discussion. The role of mining in modern power systems is one of the conversations on the agenda.

Follow Bitcoin Poland Conference

Website: bitcoin-poland.com

X: @Bitcoin_Poland

LinkedIn: Bitcoin Poland Conference

Tickets: https://tobilet.pl/bitcoin-poland-conference-2026.html