pull down to refresh
AI will go through this in its own way (but it will be a combination of software and hardware disruption)
Usually hashrate keeps grinding higher through the pain. Price falls, miners get squeezed, weaker operators sell or restructure, but the network keeps making new highs because newer machines, cheaper power, and deployment pipelines keep pushing hash online. There's also a lag between purchasing, delivering and installing of machines.
This time looked different. Just look at the shape of the curves below - it's a gradual grind down. This is different. Seems like miners are all gradually being flushed out. Great for those that keep their compass pointing north and can weather the storm.
Have to admit I'm not a heavy lightning user... But....
I've used Breez, Aqua, Wallet of Satoshi and some times Cash App. Tether Wallet also has lightning now.
Also used Zeus on my node.
I do think a lot of the mempool being empty is because of: 1. the bear market; 2. a lot of financial volume went to ETFs (separate discussion if good or bad). At least some of the trading volume below would have gone to the bitcoin blockchain.
ETFs do bring new players as well. I think it's a net win for Bitcoin in the long term. But brings noise in the ST.
Sure thing. We've open sourced most of the research analytics we use at Elektron there.
I've been a huge proponent of Open Source Software for a while. With AI it becomes inevitable (easy to reverse engineer almost anything).
Lots of data, simulations and charts there!
Absolutely. Fundamentally nothing changed with Bitcoin through this bear market. Still the best freedom money ever found by humans.
Bear markets are brutal... And we tend to think it's game over. We've been here before. And will be here again.
I wrote this in one of my newsletters:
Mining is a buyer of last resort for electrons when no one else needs them, and a seller of first resort when the grid is stressed. It is the most agile load the power sector has ever seen.
Yes, there is still a lot of opportunity.
The obvious low-hanging fruit has mostly been identified: flare gas, curtailed hydro, stranded wind and solar, remote generation with no transmission, weak grids with excess off-peak power. People know where many of these pockets are.
There's still a massive educational effort that needs to happen with energy and utility companies. They hear Bitcoin and come to all the wrong conclusions.
But we've had good success in places like Brazil where there's a ton of excess wind and solar built with limited distribution.
I truly believe the risks on AI/HPC are mostly miss priced. A few points:
- Much higher capex needs than BTC mining. The capex is much heavier per MW.
- Technology is just emerging. AI will see their version of CPU => GPU => ASICs.
- In AI you sign contracts in fiat. In BTC mining you get the best savings technology ever invented. (of course you could always buy BTC with AI revenue).
I've been in love with computers and coding since a very young age. Then I went to work in finance (Economics intrigued me because I grew up in Brazil with raging inflation). I was also lucky to learn Austrian School Economics back in 1997. Way before it became popular.
I was primed (and lucky) to grok Bitcoin early.
Section 5 below address this as well in more detail:
https://x.com/alphazeta/status/2053918448065929516
I'm convinced we're witnessing the first bear market in bitcoin's hashrate. Wrote about it in the article below. On Section 4 I discuss exactly this point:
TL;DR quotes:
First, two things that are almost always confused. Private-key security, the cryptography you rely on when you sign a transaction or move coins, has nothing to do with hashrate. Elliptic curves and SHA-256 don't care how many machines are mining. Your keys are as safe at 400 EH/s as at 1,200 EH/s; that is a different problem entirely. This nuance is also frequently missed on quantum discussions but I’m not going there on this newsletter.
What protects Bitcoin isn't the absolute number on the hashrate chart. It's the cost asymmetry between attackers and the value at risk, and that asymmetry remains enormous.
Epoch after epoch (in fiat terms) - daily miner rewards keep going up:
https://elektronics.dev/analytics/blockchain
I’m optimistic over the long run, but I think this is the most important unresolved question in Bitcoin’s hundred-year model. Anyone who says it is obviously solved is hand-waving.
Today’s fee market is not carrying the load. The subsidy still pays miners. Fees are tiny most days. After the ordinals/runes spike, fee share collapsed again, and the mempool has been quiet (lots of txs went to ETFs, and other factors collaborate here).
Long term I'm optimistic: Bitcoin base layer blockspace becomes more valuable as Bitcoin itself becomes more important. High-value settlement, self-custody, Lightning channel management, institutional cold-storage movement, all create real demand for final settlement...
I do not dismiss AI. It is real. It is a massive new source of power demand, and any serious energy company has to understand it.
But Bitcoin mining and AI are not the same business.
Finally, I think the time to invest in AI will be after this current cycle. Lots of froth right now... Let it deflate first.
Have you seen this one?
https://elektronics.dev/analytics/pool-analysis