Bitcoin Mining’s Longest Downturn Faces a 19% Return Test

Bitcoin’s current mining slowdown is the longest in the network’s history, according to Saifedean Ammous, who estimates miners need almost 19% annual appreciation over four years to sustain economics after 50% subsidy cuts.

In a mining update the author of The Bitcoin Standard recently reshared, Ammous said: “the current period has become the longest in Bitcoin’s history.” He cited Bitcoin’s earlier price collapse of roughly 94% to 95% and the exceptionally long difficulty downturn that followed China’s 2021 ban.

Ammous attributed the slowdown in part to structural halving economics, not to a shock comparable with China’s 2021 mining ban. His account says China’s sweeping ban on domestic crypto mining forced a substantial portion of global hash rate to relocate, while the present slowdown has occurred without anything remotely comparable.

What changed in Bitcoin’s mining economics?

Bitcoin’s pre-programmed halvings reduce the block subsidy by 50% roughly every four years. Miners then need BTC’s price to rise fast enough to offset lower rewards. Ammous estimated that mining economics require the cryptocurrency’s value to appreciate by almost 19% annually over four years, before accounting for dollar devaluation and the growing opportunity cost created by AI infrastructure.

The 19% figure is a four-year operating hurdle. Ammous’s calculation combines scheduled subsidy reductions, dollar devaluation and competition from AI infrastructure for capital and electricity. If expected returns fall below the threshold, allocating more capital and electricity toward mining becomes progressively less attractive.

For miners, that hurdle changes the case for every marginal deployment of rigs, facilities and electricity. When AI infrastructure can use the same power and data-center footprint to generate colocation revenue, mining expansion must clear a higher bar than an alternative use of the same capital.

Ammous also argued: “Bitcoin miners no longer have the same monopoly over cheap power.” In his account, AI and high-performance computing operators now have a stronger claim on electricity and data-center capacity, making colocation a direct operating alternative for sites able to serve both workloads.

Why are miners choosing AI colocation?

Core Scientific said it plans to fully exit Bitcoin mining by the end of the year and enter 2027 as a dedicated AI and high-performance computing colocation provider. It has already reduced its mining fleet by almost half as it redirects electricity and data-center capacity toward infrastructure customers.

Core Scientific’s colocation revenue jumped 58% quarter over quarter to $215.4 million in the third quarter, compared with $15 million a year earlier. The increase gives the company a larger revenue base outside mining and makes converting power and data-center space into infrastructure services more immediate than adding mining capacity.

That shift does not eliminate pressure on miners without alternative customers for their power. Under Ammous’s framework, they still face a 50% subsidy reduction, an estimated 19% annual appreciation requirement and competition for investment from AI infrastructure. Operators able to secure colocation demand can redirect existing power and facilities; miners without that demand face lower expected returns.

Could a mining slowdown support Bitcoin holdings?

Ammous does not necessarily view the contraction as bearish for BTC. His argument is that mining’s declining appeal could redirect capital rather than remove it: money that might have financed machines, facilities and electricity could instead flow directly into Bitcoin. He also expects declining electricity consumption to weaken one of the network’s persistent criticisms, namely its energy use.

The broader consequence is a reordering of investment priorities among rigs, facilities, electricity contracts and data-center capacity. It does not erase mining risk; it widens the gap between operators with infrastructure customers and miners exposed only to Bitcoin economics.

The next concrete checkpoint is year-end, when Core Scientific’s planned mining exit takes effect, followed by its entry into 2027 as an AI and high-performance computing colocation provider. Its next revenue disclosures will test whether infrastructure can replace mining at the scale implied by the 58% quarterly increase in colocation revenue.