BTC is trading at $62K.

JPMorgan estimates the all-in production cost at roughly $78K.
Bitcoin has now traded below that level for five consecutive months.
The last time something similar happened was 2018.
Not because the prices match.
Because the miner behavior does.
Back then, the sequence looked like this:
> BTC traded below production cost for months
> Miners sold aggressively to fund operations
> Difficulty adjusted lower multiple times
> Marginal operators exited
> Forced selling pressure faded
> BTC eventually recovered
We’re seeing several of those conditions again.
Public miners sold more than 32,000 BTC in Q1 2026 alone.
Mining difficulty has dropped by more than 10% twice this year.
JPMorgan estimates $BTC remains roughly 19% below production cost.
The pattern rhymes.
But the structure is different.
In 2018, mining companies had one business: Mine Bitcoin.
In 2026, many have two.
MARA, Riot, and others are increasingly monetizing the same infrastructure through AI and HPC hosting agreements.
That changes the equation.
Operators can remain online longer even when mining economics deteriorate.
The second difference is the buyer base.
In 2018, miner selling largely met retail demand.
Today, ETFs, corporate treasuries, and institutional allocators absorb a meaningful share of market supply.
The third difference is hardware economics.
ASIC fleets represent billions of dollars of sunk capital.
Many operators are more likely to continue mining at a loss and sell BTC to fund electricity than shut machines down entirely.
That slows the capitulation process.
Which brings us to the key question.
Are we seeing a 2018-style miner capitulation?
Not quite.
The miner liquidation phase appears advanced.
The difficulty adjustments have started.
But the structural support mechanisms didn’t exist in 2018.
AI revenue extends survival.
Institutional buyers absorb supply.
Larger hashrate bases slow the cleansing process.
…
My take
The 2018 analog is useful because the mechanisms are similar.
Below-cost mining eventually forces supply-side adjustment.
The difference is that 2026 has more buffers between miner stress and miner shutdown.
The pattern may rhyme.
The timeline probably doesn’t.