L7MLb (1).jpg


Imagine you spent years building a massive factory that prints gold. Then a new neighbor shows up and offers to pay three times more for the exact same electricity — with long-term fixed contracts. What do you do? You retool the factory.

That’s exactly what’s happening with large publicly traded Bitcoin miners right now.

In the first half of 2026, a group of 12 major mining companies wrote off roughly $1.1 billion in asset value. Most of it was mining equipment and infrastructure they no longer plan to use for Bitcoin. The hashrate they removed (about 75 EH/s) is estimated to represent another $1.5 billion in hardware at acquisition cost. Just two companies — IREN and Core Scientific — accounted for nearly 90% of those write-downs.

Why the sudden shift? After the 2024 halving and the subsequent drop in Bitcoin’s price, mining economics turned ugly. The average cash cost to produce one Bitcoin for public miners climbed above $75,000 in some quarters, occasionally touching $80,000. Meanwhile, AI and high-performance computing (HPC) deliver roughly three times the profit per megawatt — and, crucially, that revenue is predictable. Companies are locking in 10- to 20-year contracts with Microsoft, CoreWeave, Anthropic, Google-backed firms, and others.

So miners are doing what any rational business would do:

  • Writing off or selling older ASICs
  • Paying cancellation fees to exit new equipment orders (Core Scientific paid nearly $42 million to walk away from 15 EH/s of next-gen machines)
  • Converting sites to GPU-ready AI data centers
  • In some cases, exiting Bitcoin mining entirely (Keel, formerly Bitfarms, has already stopped; IREN aims to be largely done by the end of 2026)
As a result, network hashrate has been declining. Public miners have shed a noticeable share of capacity, and difficulty has adjusted downward — in some periods posting its first year-over-year drop in a long time.

What This Means for Bitcoin Holders​

The upside:

  • Less competition → lower difficulty → higher profitability for the miners who stay. This is Bitcoin’s classic self-regulating mechanism at work.
  • The companies that remain will be stronger and more efficient.
  • Over the longer term, a leaner, more professional mining sector can actually improve network health.
The downside and risks:

  • While the transition is underway, hashrate is softer. Network security is slightly reduced in the short term (though still far from any critical level).
  • Miners are selling portions of their Bitcoin treasuries to fund AI data-center builds, adding supply to the market.
  • If too many large players leave permanently and new capacity doesn’t replace them quickly enough, the network’s security budget (miner rewards) could grow more slowly than ideal.
Interestingly, most analysts believe that even a strong Bitcoin price recovery is unlikely to pull the majority of these companies back into mining. The AI contracts are long-term, the infrastructure has been reconfigured for GPUs, and the margins are simply better. Reversing the pivot would be expensive and unattractive.

This isn’t the death of Bitcoin mining. It’s a maturation and a split. Some companies are becoming AI infrastructure players (and the market is rewarding them for it — many stocks have rallied hard on contract announcements). Others are doubling down as pure-play miners in a tougher, cleaner competitive environment.

For the average holder, this is a neutral-to-positive development over the medium term. Bitcoin’s network is designed to absorb structural shifts like this. Hashrate has crashed before (China 2021, post-halving purges) and has always recovered at a new equilibrium.

The clear signal: the era of “hashrate growth at any cost” is over. Economics now rules. And as long as AI pays better, miners will choose AI.
  • Like
Reactions: Hunter_245