Stories · as read on the Mining station
Core Scientific's Russell Cann on Hashprice and Power
Machines now last longer, hashprice swings harder and AI bids up power. Cann's case for curtailment, hedging and planning for 2028 and 2032.
Core Scientific builds a Bitcoin mining site for about half a million dollars a megawatt. Converting one for AI, Russell Cann told a room in Miami, runs roughly eight and a half to nine million per megawatt.
Cann is Chief Customer Success Officer at Core Scientific. He spoke at Mining Disrupt Miami, June 24 to 26, 2024, on a topic he warned would be dry: hashprice volatility, and why it matters.
He started with hardware. In past cycles, he said, chips got much more efficient at the same time as each halving, so miners came to believe every halving meant scrapping the whole fleet. He called that a coincidence. The gap between new generations is now small, so machines last longer. Even Core's 34-joule machines are still profitable on many days, and when they are not, they get turned off.
His second point was about what the company watches. In his words: "There's only one person at Core that gives a shit about the price of BTC." That is the trader. Everyone else watches hashprice, which takes in both network difficulty and the Bitcoin price. On the day of the talk, it sat just below 5 cents.
Then power. In the United States, he said, AI buyers are pushing power prices up, and they can pay far more than miners. Core's all-in price was about 4.25 to 4.5 cents. He wants closer to 2 cents for mining, with the option to switch off when hashprice drops. AI buyers, he expects, will offer utilities 6 or 7 cents.
That is why Core had set aside 700 of its 1,200 megawatts for AI. He said it was not an exit from mining but a move to the highest and best use, with new and more efficient mining sites to follow.
He was clear that the move is not for everyone. On top of the conversion cost, filling a site with H100s adds about $25 million a megawatt. And running mining and AI on the same power contract was, in his words, not just no, but hell no.
His advice to new miners was to plan for the 2028 and 2032 halvings, when transaction fees will have to carry far more of the load, and to hedge both ends: power going in, production coming out, the way oil companies did in the eighties. As for running at a loss, he said, "Mining at a loss is just stupid."
When Cann spoke, Bitcoin traded at $61,272. Today it is $81,466.
The full story is on Mining Disrupt at miningdisrupt dot com, and many more like it play around the clock on Bitcoin Radio.