Bitcoin Miners Sell BTC Reserves as Hashrate Crunch Crushes Margins

Bitcoin mining is turning into a capital-sucking meat grinder, and some of the biggest public miners are responding by selling Bitcoin reserves, chasing AI revenue, and scrambling for more flexible ways to stay alive.

• Network hashrate remains near record highs, crushing mining profitability
• Hashprice has collapsed, making each unit of mining power far less valuable
• MARA and other public miners are tapping BTC treasuries to fund operations and AI pivots
• Multi-coin mining is emerging as a more agile alternative to Bitcoin-only fleets

The economics behind Bitcoin mining have gone from competitive to downright vicious. Late in 2025, the Bitcoin network reportedly hit 1.25 zh/s in hashrate, and even in H1 2026 it remained extremely elevated at 958.01 EH/s. For readers who don’t live and breathe mining charts, hashrate is the amount of computing power securing the network. Higher hashrate generally means more competition, harder mining, and thinner margins.

That pressure shows up most clearly in hashprice, which is basically the revenue a miner earns per unit of hashpower. According to Hashrate Index data, hashprice has dropped by more than 90% since 2021, falling from $0.400 per TH/s/day to just $0.035 per TH/s/day. That is not a gentle decline. That is a business model getting kneecapped in broad daylight.

The estimated total cost to produce one Bitcoin has also climbed to around $86,944, a figure that helps explain why so many miners are under strain. Mining costs are not just electricity bills. They include hardware purchases, facility costs, maintenance, staffing, financing, and the brutal reality of depreciation. A miner buys expensive machines that start losing value almost immediately, then has to keep them fed with cheap power in a network that keeps getting harder to compete against. It’s a lovely little stress test for balance sheets. If your spreadsheet looks even a bit weak, Bitcoin mining will find out fast.

Hardware pricing tells the same story. A new Bitmain Antminer S19 is reportedly priced around $2,511, while used units are being dumped for as little as $99 on secondary markets. When yesterday’s industrial-grade money printer gets treated like a garage sale impulse buy, you know the market is not exactly thrilled with mining margins.

That squeeze is forcing a painful shift among public Bitcoin miners. Instead of behaving like long-term BTC accumulators, some are behaving like distressed capital allocators. MARA Holdings reportedly held 53,822 BTC in late March 2026, but by May 19 that treasury had dropped to 35,303 BTC. The figures cited suggest MARA liquidated around $1.5 billion worth of Bitcoin in one quarter to cover operations and fund a pivot into AI digital infrastructure.

That move makes sense if you strip away the marketing gloss. “AI digital infrastructure” usually means data centers, compute capacity, and the power and cooling systems needed to host modern AI workloads. In other words, the same steel-and-electricity skeleton that once supported Bitcoin mining can sometimes be repurposed for a different customer base. If mining profits are weak, AI hosting can look a lot more attractive. Capital goes where the margin is, not where the slogan is.

Still, there’s a hard truth buried in that pivot: if a miner has to sell a large chunk of its BTC treasury to fund the next phase of the business, it stops looking like a pure Bitcoin accumulation vehicle and starts looking like a leveraged infrastructure company with a crypto habit. Not exactly the clean “mine and hold” fairy tale many investors imagined when public miners became a popular stock proxy for Bitcoin exposure.

That’s why the concept of rigid infrastructure has become such a problem. If a mining operation is locked into one coin, one algorithm, and one network’s economics, it has very little room to maneuver when profitability gets squeezed. Difficulty adjusts, hashprice falls, and suddenly the fleet becomes a liability instead of an asset. The operations that survive aren’t necessarily the ones with the biggest pile of machines. They’re the ones with the best optionality.

That is where more flexible, multi-network mining models come in. HashNet says it has a $300M+ global footprint across six cryptocurrencies and four algorithms, and claims its “Alpha Engine” can switch algorithms in 12 milliseconds. The idea is simple: point hashpower at whichever proof-of-work network is paying best right now, then convert those gains back into Bitcoin. It is the opposite of being trapped in one lane while the rest of traffic is taking different exits.

“Our Alpha Engine switches algorithms in 12 milliseconds to capture the highest return.”

That approach is especially relevant when altcoin mining opportunities flare up. The piece points to Zcash as an example, citing a 1,900% rally between September and November after a network upgrade, followed by another 119.5% run in early 2026 ahead of a further upgrade. Those kinds of moves can make opportunistic mining look brilliant. If a network upgrade drives price higher and mining economics improve at the same time, flexible operators can rack up outsized returns while Bitcoin-only fleets sit there grinding away for thinner and thinner margins. The Hardware Trap: Why Bitcoin Miners Are Liquidating the Asset They Built

That does not mean altcoin mining is some magical escape hatch. It comes with obvious trade-offs: smaller markets, higher volatility, sharper drawdowns, and the constant risk that an asset’s economics or community support can evaporate faster than a leveraged trader’s confidence on a Sunday night. Flexibility is useful, but flexibility alone does not make a business durable. It just makes the business less dumb.

There’s also a strategic argument for the Bitcoin-only crowd that should not be ignored. Bitcoin remains the most valuable and credible proof-of-work asset by a mile. Its monetary premium is the point. A miner that stays aligned with Bitcoin may avoid chasing junky yield in marginal networks that can blow up just as fast as they pump. For some operators, the cleanest strategy is still to secure BTC, hold BTC, and avoid getting cute.

But the market has a nasty habit of rewarding survival over ideology. Public Bitcoin miners are not hobbyists, and they are not protected by conviction alone. They answer to shareholders, lenders, and operating costs. When mining profitability collapses, the treasury becomes a war chest, not a sacred cow. And if a company can earn a better return repurposing its infrastructure for AI workloads, that capital will move. No one gets points for purity when the electricity bill is due.

“Bitcoin production costs spike to $86,944 as network hashrate maintains massive overhead pressure at 958 EH/s.”

“Corporate giants like MARA liquidate billions in BTC reserves to fund expensive infrastructure pivots into AI.”

“The future of mining may not belong to the largest single-coin fleet, but to the most flexible one.”

That last line captures the real shift underway. Bitcoin mining is no longer just about stacking as much hashpower as possible and waiting for the price to save you. It is becoming a capital efficiency contest. Energy contracts matter. Hardware depreciation matters. Treasury management matters. Optionality matters. A miner that can rotate between proof-of-work opportunities, monetize infrastructure in other ways, and avoid sitting trapped in a bad margin environment has a much better chance of surviving long enough to accumulate BTC.

Why are Bitcoin miners selling BTC?
Because mining margins are under severe pressure, and public miners need liquidity to pay operating costs, service financing, and fund new infrastructure.

What is hashprice?
Hashprice is the revenue a miner earns per unit of hashing power. When it falls, each machine generates less income, which crushes profitability.

Why does mining difficulty matter?
Mining difficulty is the network’s way of adjusting how hard it is to find blocks. When more hashpower enters the network, difficulty rises and competition gets tougher.

Why are miners pivoting into AI data centers?
Because the same physical infrastructure used for mining can sometimes be repurposed for AI compute hosting, which may offer steadier revenue.

Is Bitcoin-only mining a bad idea?
Not necessarily. It can still make sense for operators with strong power access and disciplined balance sheets, but it leaves little room to adapt when margins collapse.

Does flexible multi-coin mining guarantee better returns?
No. It adds complexity and exposure to smaller, more volatile assets. It can improve resilience, but it is not a free lunch.

The uncomfortable takeaway is simple: the biggest Bitcoin mining companies may not survive by acting like Bitcoin purists. They may survive by becoming ruthless about cash flow, ruthless about infrastructure reuse, and ruthless about exiting bad positions before the market does it for them. That is not romantic, but neither is watching your BTC treasury get liquidated because your mining fleet turned into an expensive heater with branding.

Bitcoin still wins as the monetary asset. No question. But Bitcoin mining companies are increasingly being forced to behave like energy and compute businesses first, and Bitcoin accumulation machines second. The survivors will be the ones that can move fast, cut losses, and build for flexibility instead of worshipping hardware like it’s a shrine.