Bitcoin miners are bleeding at $90,000, but the “death spiral” math hits a hard ceiling

Bitcoin’s “miners are dumping” story is comforting in the way simple stories always are. Price slides, miners run out of oxygen, coins hit exchanges, and the price is shoved around by a single, easy villain.

But miners are not a single actor, and selling pressure isn’t just a mood. It’s math, contracts, and deadlines. When stress shows up, what matters is not whether miners want to sell, but whether they have to, and how much they can sell without breaking the business they’re trying to keep alive.

That’s why the most useful way to think about a miner “capitulation” is as a thought experiment. Imagine you’re running a mine right now, in a market where the hashrate ribbon flipped into inversion territory, and price trades below a rough, difficulty-based estimate for average all-in sustaining cost, around $90,000.

At the same time, total miner holdings sit at around 50,000 BTC: not small by any measure, but not bottomless either.

Now you’ve got a simple question that sounds dramatic. If price sits below the average AISC line for a while, how many coins can you push out over 30 to 90 days before lenders, power contracts, and your own operating reality push back?

AISC is a moving target, not a single number

All-in sustaining cost, or AISC, is crypto’s borrowed term from mining and commodities, but it earns its keep because it forces you to stop pretending electricity is the only bill. AISC is basically a number that determines whether you can stay in business. Not “can you keep the machines on today,” but “can you keep the operation healthy enough that it still exists next quarter.”

You can think of Bitcoin miners’ AISC as having three layers, even if different research shops draw the boundaries differently.

The first layer is the one everyone understands: direct operating cash costs. Electricity sits at the center of it, because the meter runs whether you’re feeling bullish or not. Add hosting fees (if you don’t own your site), repairs, pool fees, network ops, and the people who keep the facility from turning into an expensive space heater.

The second layer is the one the memes skip: sustaining capex. This isn’t growth capex: sustaining capex is the money you spend to stop your fleet from slowly dying. Fans fail, hashboards degrade, containers rust, and, more importantly, the network gets tougher. Even if your machines are fine, you can lose a share of the pie if everyone else upgrades and you don’t.

That’s where difficulty comes in. Bitcoin adjusts mining difficulty so blocks keep arriving roughly on schedule. When hashrate rises, difficulty ratchets up, and the same machine earns fewer BTC for the same energy burn.

When hashrate falls, difficulty can ease, and the remaining miners get a slightly better bite. The AISC framing we’re using is explicitly based on difficulty, which is a clean way to capture this moving target without needing every miner’s private power contract.

The third layer is what turns stress into forced behavior: corporate costs and financing. A private operator might care mostly about power and maintenance. A public miner with debt cares about interest payments, covenants, liquidity buffers, and the ability to refinance.

This is why AISC changes over time in a way that makes single-number debates feel silly. It changes when difficulty changes, and when the fleet mix changes (older machines get pushed out, newer ones come in).

It changes when the power environment changes, especially for miners exposed to spot pricing, and it changes when capital costs change, which is why a miner can look stable at one point in the cycle and fragile at another with the same hash output.

So when price dips below an average AISC estimate like ~$90,000, it doesn’t mean the whole network is instantly underwater, just that the center of mass is uncomfortable. Some miners are fine, some are pinched, and some are in triage. The stress is real, but the response is uneven, and that unevenness is what keeps the “everyone dumps at once” from being the default outcome.

There’s another reason the default outcome isn’t a dump. Miners have more levers than just selling their BTC: they can shut down marginal machines, curtail for grid payments, roll hedges, and renegotiate hosting terms. And, as previously covered by CryptoSlate, many now have side businesses tied to AI data-centers, which can buffer a bad mining month.

That gets us to the real question, which is when stress is on, how much selling is structurally required?

The dump math: what can be sold without breaking the business

Start with the one flow the protocol hands you, whether you’re happy about it or not. Post-halving, new BTC issuance from the block subsidy is about 450 BTC per day, which is about 13,500 BTC per month.

If miners sold 100% of new issuance, that’s the clean ceiling for flow selling. In reality, miners don’t coordinate, and not all of them need to sell everything they mine. But as a thought experiment, 450 BTC/day is the maximum new supply that can hit the market without touching any pre-existing inventory.

Now bring in inventory, because that’s what the scary headlines point at. We’ll rely on Glassnode’s estimate that miners have around 50,000 BTC on hand. A 50,000 BTC stockpile sounds large until you turn it into a time series. Spread across 60 days, 10% of that inventory is 5,000 BTC, which is about 83 BTC/day. Spread across 90 days, 30% is 15,000 BTC, which is about 167 BTC/day.

That’s the basic shape of miner forced distribution in a stress window: flow selling does most of the work, and inventory selling adds a smaller but still meaningful amount, unless the stress is severe enough that inventory becomes the primary tool.

So let’s put three price paths under the toy model: $90,000, $80,000, $70,000. Then tie them to three middle-ground regimes that map to how miners behave when margins get thin.

In the base case, miners sell half of the issuance and touch no inventory. That’s 225 BTC/day. Over 60 days, that’s 13,500 BTC of issuance in total times 50%, so 6,750 BTC. Over 90 days, 10,125 BTC.
In a conservative stress case, miners sell 100% of issuance and also sell 10% of inventory over 60 days. That’s 450 BTC/day from issuance plus 83 BTC/day from inventory, about 533 BTC/day total.

In a severe stress case, miners sell 100% of issuance and sell 30% of inventory over 90 days. That’s 450 plus 167, about 617 BTC/day.

Those are the upper-bound sketches that answer a narrower question: what does the market allow?

To understand how much the market would notice, we’ll use the simplest comparator readers already understand: ETF flow days, measured in BTC-equivalent. ETF outflows are only around 2.5% of BTC-denominated AUM, about $4.5 billion, and CryptoSlate previously described them as more technical than conviction-driven. You don’t even need to litigate motive to use the comparison, because the point is scale.

At $90,000 per coin, a $100 million day is about 1,111 BTC. At $80,000, it’s 1,250 BTC. At $70,000, it’s about 1,429 BTC. Suddenly, the miner numbers look less like a monster under the bed and more like something you can place on the same shelf as flows the market digests all the time.

A severe miner distribution sketch, say 600 BTC/day, is roughly half of a $100 million ETF day at $90,000. That can still move price if it’s dumped into thin books, or if liquidity is fragile on a weekend, or if it clusters into a few ugly hours. But the brute-force story of miners flooding the market runs into two ceilings: the issuance and the finite inventory that miners are willing and able to liquidate.

There’s also the execution detail that matters more than people want it to. A lot of miner selling is not a market order slapped into the public order book. It can be routed through OTC desks, structured as forward sales, or handled as part of broader treasury management. That doesn’t erase selling pressure, but it changes how it prints on the tape. When people expect a visible waterfall and get a slow grind, the effect on the market is dampened.

So what would turn this from an orderly drip into something uglier? It would certainly require more than just the price dropping below ASIC. The trigger is when the financing layer takes over the decision. If a miner needs to defend a liquidity minimum, meet collateral terms, or handle a refinancing wall in bad market conditions, then inventory turns from optional to necessary.

That’s the sober answer to the viral question. Even when stress is on, and the ribbon is inverted, there are real limits to what miners can dump in a month or a quarter. If you want a practical ceiling, the thought experiment keeps pulling you back to the same zone: a few hundred BTC per day in mild stress, and something like 500 to 650 BTC per day in harsh stress windows that include inventory taps, with the exact number hinging on power terms and debt constraints you can plug in later.

And if you’re trying to guess what moves the tape, the punchline is annoyingly unromantic. The market tends to care less about the narrative label on a seller and more about the cadence, the venue, and the surrounding liquidity. Miners can add weight to a down week, but the idea that they have an infinite trapdoor under price does not survive contact with the balance sheet.