Because the bitcoin gold rush dries up, crypto miners are finding it hard to make ends meet. But for a lot of there’s a silver lining—the facilities they’ve arrange are perfect for Silicon Valley’s latest obsession with artificial intelligence.
Crypto mining could be a profitable but highly volatile endeavor. It involves creating massive datacenters full of specialized computer chips and using them to unravel the mathematical puzzles underpinning the safety of varied cryptocurrencies. In exchange, the miners win a few of that cryptocurrency as a reward.
Most miners make the majority of their money from bitcoin. But earlier this yr, an event called “the halving” seriously hit earnings. Every 4 years, the bitcoin protocol halves the mining reward—that’s, how much bitcoin miners receive in exchange for solving math puzzles—to extend the scarcity of the coin. Normally, this causes the worth of bitcoin to leap in response, but this time around that didn’t occur, severely impacting the profitability of miners.
Fortunately for them, one other industry with a voracious appetite for computing has arrived just in time. The push to coach massive generative AI models has left firms scrabbling for chips, datacenter space, and reliable access to large amounts of low-cost power, things many miners have already got in abundance.
“It [normally] takes 3-5 years to construct an HPC-grade data center from scratch,” JPMorgan analysts wrote in a recent note, in response to the Financial Times. “This scramble for power puts a premium on firms with access to low-cost power today.”
While crypto mining and training AI aren’t the exact same, they share crucial similarities. Each require huge datacenters specialized to perform one particular job, and so they each devour large amounts of power. But because miners have been playing this game for a very long time and most AI firms have only began attempting to train truly massive models because the launch of ChatGPT lower than two years ago, the businesses have an enormous head start.
They’ve already spent years scouring the country for places with abundant low-cost power and lots of space to construct large datacenters. More importantly, they’ve already undergone the time-consuming strategy of getting approvals, negotiating power licenses, and getting the facilities up and running.
The rapid expansion in demand for AI training is straining grids in some areas, and so, many jurisdictions in North America have implemented long waitlists for brand new datacenters, in response to Time. Already, roughly 83 percent of datacenter capability currently under construction has been leased prematurely, says Bloomberg.
This implies the most important bottleneck for a lot of AI firms is finding the hardware to coach their models, and that presents a brand new opportunity for crypto miners. “You’ve seen plenty of crypto miners that were form of struggling which have actually made a full pivot away,” Kent Draper, chief business officer of crypto miner IREN, told Time.
Converting a bitcoin mine into an AI training cluster isn’t a straight swap. AI training is usually done on GPUs while bitcoin mining uses specialized mining chips from Bitmain. But often, it’s not a lot the chips AI firms are after, however the infrastructure and power access the mine has already arrange.
In June, crypto miner Core Scientific announced it might host 270 megawatts of GPUs for the AI infrastructure startup CoreWeave. “We view the chance in AI today to be one where we will convert existing infrastructure we own to host clients who want to install very large arrays of GPUs for his or her clients which might be ultimately AI clients,” Core Scientific CEO Adam Sullivan told Bloomberg.
Some miners are also operating GPUs themselves. German miner Northern Data had already purchased $800 million of Nvidia GPUs for mining the Ethereum cryptocurrency, but a serious software update to the coin’s blockchain in 2022 did away with mining and meant those chips were sitting idle. The corporate has now repurposed them right into a 20,000-GPU training cluster, one among the most important in Europe, in response to Bloomberg.
Other miners like Hut 8 and IREN are investing heavily in recent chips to more proactively chase the AI boom. Often, AI training is going on side-by-side with crypto mining. “We view them as mutually complementary,” IREN’s Draper told Time. “Bitcoin is fast revenue but somewhat more volatile. AI is customer-dependent—but once you might have customers, it’s contracted and more stable.”
This recent trend could provide some modest environmental advantages too. Persons are concerned concerning the enormous power consumption of each AI training and bitcoin mining. If increasing demand for AI simply displaces existing mining infrastructure, slightly than requiring recent power-hungry datacenters, that might help curtail the growing carbon impact of the industry.
Nonetheless, for miners, chasing the most recent gold rush could be a dangerous strategy. There are growing concerns the AI industry is in a bubble near bursting. If that happens, the wealthy recent seam miners have began to tap could dry up in a short time.
