This is a reasonable, non-breathless take on the state of bitcoin mining from Elektron Energy CEO, Ralph Zagury. Lots of good data and analysis here and a tease of a mining data website they are building.

Estimated network hashrate set its all-time high at 1,275 EH/s on September 19, 2025. It now sits in the 900 EH/s range, which works out to a drawdown of somewhere around 22% to 24% depending on methodology. On our own measurement, this is now approaching the longest stretch in the past decade without a new ATH, and the trend is not flat or consolidating; it is quietly grinding lower. So, this isn't a mirror image of 2021.

I knew that there had been some negative difficulty adjustments, but I didn't realize it was quite to this level.Down ~20% is surprising to me.

Bitcoin's supply is perfectly inelastic. It doesn't care that price is high, or that ASICs got 30% more efficient last year, or that you happened to buy a hundred thousand of them. Blocks arrive every ten minutes on average, and they will keep arriving every ten minutes on average forever.

The flip side is what we’ve been witnessing recently. When miners leave, the network adjusts in favor of the ones who stay: difficulty falls, and the lean operators end up with more blocks per unit of hashrate. A bigger slice of a fixed pie.

Bitcoin mining is so crazy. It feels like people will be puzzling over it for the next hundred years. We just aren't used to things that work like this.

Fees as a percentage of total miner rewards have collapsed since the 2024 ordinals/runes peak. Most days, fees contribute less than 0.5% of block rewards, and months can pass without a meaningful spike.

Why isn't fee revenue growing? Zagury speculates that the ETFs and other financialized versions of bitcoin took a lot of the transaction volume.

ETFs really do function like an institutional abstraction layer for Bitcoin exposure. Secondary-market trading happens in securities markets (NYSE, NASDAQ) and doesn't require every buyer and seller to touch the chain. Most institutional Bitcoin sits in cold custody, settles in fund accounting, and only broadcasts on- chain on creation or redemption. That's good for adoption, but it may also reduce the on-chain footprint per dollar of exposure, which is exactly what the fee market eventually needs to inherit.

Of course, just because fees are done and halvings happen every four years, it doesn't the mining industry is making less money: in dollar terms, it's never been bigger:

In dollar terms, aggregate miner revenues have generally increased from epoch to epoch despite each halving. Era 3 averaged roughly $2.1 million per day, Era 4 roughly $11.6 million, Era 5 roughly $31.9 million, and Era 6 so far roughly $40.1 million. That is the power of Bitcoin’s price appreciation. But this is the network-level story, not the individual miner story. The total pie can grow while each miner’s slice gets smaller. As estimated network hashrate rises, a fixed fleet controls a smaller share of the network and has a lower probability of finding blocks. So even if total miner revenue in dollars is higher than in prior epochs, hashprice can still compress and individual miners can still feel poorer. This is why aggregate miner revenue is interesting, but hashprice is the metric that tells you whether a miner is actually getting paid for each unit of hashrate.

But Zagury sees some silver lining on these clouds. First, in the heat-punk, load-balancing, home hashing movements:

None of it [Bitaxe units sitting on desks, apartment-sized hydro miners, industrial facilities embedding hashboards into existing thermal loads] is competitive with utility-scale operations today, and none of it needs to be. The point is that the economic floor for owning hashrate keeps falling, while the marginal value of putting waste heat to work keeps rising. Eventually, the hashrate distribution starts to look less like a handful of cathedrals and more like a long tail of small, distributed, semi-residential uses.

Is Zagury feeling down on the state of mining? I don't think so:

So: a hashrate bear market is here. What, is soon to be, the longest stretch without a new ATH is here. The halving is roughly two years out, and at flat prices it will wash out a meaningful piece of the network. The fee market is quieter than Bitcoin's long-term security model eventually requires. The publicly traded miners are pivoting away from us. And the mood, as I wrote in "The Quiet," is exhausted.

And yet, every one of those variables is exactly what the difficulty adjustment was built to absorb. Less hashrate means easier blocks. A bigger share for the lean means better unit economics. Weaker miners turning off means stronger miners surviving. Rebalancing.

That's the part I keep coming back to. The incentives in Bitcoin don't require anyone to be wise; they only require people to be self-interested. Wisdom is optional, discipline is rewarded, patience accrues compounding returns, and the protocol does the rest.

And finally, he concludes by unveiling this cool website they've been working on:

One last sneak peek for readers of this newsletter: over the last few months I have been quietly aggregating much of the research and analytics behind these letters into Elektronics.dev. It is still very much in beta, use at own risk. We are building it as an open analytics layer for Bitcoin mining: network data, ASIC economics, hashprice, Bitcoin price, simulations, fleet profitability, and some tools we use internally to make sense of this business. The plan is to gradually open-source the analytics and make the work auditable, reusable, and useful to anyone trying to understand Bitcoin mining beyond the headlines. More soon. Play around and please share feedback (and bugs).

I only poked around on it a little, but there is a lot there.