Anyone who follows the public Bitcoin mining markets is aware that artificial intelligence (AI) and the move to high-performance computing (HPC) have become major trends among Bitcoin miners. What started as a slow trend last year has suddenly become a strategy that many of the leading public Bitcoin miners are exploring.
This article was originally published on Blockspace Media, the authoritative Bitcoin industry publication covering Bitcoin technology, markets, mining, and ordinals. Get Blockspace articles delivered straight to your inbox by clicking here.
Core Scientific, Bit Digital, Hut 8, Hive, and IREN already have profitable AI/HPC ventures, while Crusoe Energy, Lancium, Cipher, Terawulf, Riot, and Bitfarms are in the development or research stages. With SoftBank, OpenAI, and others pledging to collectively invest up to $500 billion in U.S. AI development through the Stargate project announced in January, where will the digital oil rush of pure Bitcoin miners lead?
Kevin Dede, managing director of equity research at investment bank HC Wainwright, believes there is plenty of room for both. In a recent episode of Mining Pod's Bitcoin Stock show, Dede noted that while he wouldn't bet against miners who take AI/HPC seriously, he also wouldn't underestimate the prospects of miners who focus solely on it.
Will the launch of Stargate change the conversation around AI development opportunities for Bitcoin miners?
I think the conversation changed when Core Scientific announced the CoreWeave deal six to eight months ago. That really changed the dynamic. Another thing that maybe people don’t take into account is that Bitcoin miners can compete at different levels. The Stargate project is about hyperscale facilities, but there are opportunities to do it at smaller scales.
BitDigital and Applied Digital have demonstrated that you don’t need hyperscale to succeed. There are many customers who need access to computing, and not all of them are hyperscalers.
Riot recently made the decision to pause their 600MW Corsicana Phase 2 to evaluate it for AI/HPC. Why do you think they did that?
Riot has been forced to buy shares by activist investors, which has had a positive effect on the stock price. The company has always been resolute in its intention to continue mining Bitcoin. At an analyst meeting last June, CEO Jason Les said they had no plans to get into HPC.
Riot’s facility in Corsicana is simply spectacular. The question is: is 600 MW of HPC worth more than 600 MW of Bitcoin mining? I think the answer is yes. Demand for HPC is growing and applications are evolving; we’re just getting started. The real market is the enterprise sector, where companies are using AI to optimize processes.
Comparing Bit Digital and Core Scientific, which company's strategy do you think has the most potential?
Let's start with BitDigital. They purchased GPUs and rented space in northern Iceland to meet the needs of one customer, who I assume is somewhere in Europe, to run models. Now, Iceland and Europe are not as close as you might think, which is important if they are running inference computations, as mainland Europe would be the main customer for that.
The Enovum deal went through and they got their first site, which is about four megawatts. They also just opened another site that they hope to power by this summer, initially targeting five megawatts with plans to expand to 35 megawatts of HPC capacity this year. Sam Thapar, their CEO, often points out that this acquisition opened the door to a potential 288 megawatts of HPC capacity.
When it comes to assessing risk, it comes down to a number of factors. Bit Digital acquired a company with a proven track record of building and operating these types of facilities. But of course, that adds another layer of risk on top of the base execution risk. You layer on top of that the risk that they might run into problems building and operating the next set of facilities as the year progresses.
As for Core Scientific, I would be the last person to underestimate them. They have attracted some really impressive talent. I asked their CEO Adam Sullivan about how all of their plans are coming together. He said that in the existing data center world, there are a lot of people whose career paths have limited room for growth. So if you are an employee at one of these companies and you get an offer from Core Scientific with stock options, you think, “My current options are in the double digits, but this could go up to the high double digits or even triple digits.” That’s how they have managed to attract such great talent.
On the other hand, the new B200 chips they are using are much more powerful, but also much more complex, and that could play a role in the delays in CoreWeave deployments at Core Scientific sites. I think a lot of this will come out in Core Scientific's next earnings call in March. They will likely discuss whether they are still on track to launch the first major CoreWeave site in Q2, and how they have dealt with these networking issues.
Do you think the higher potential of AI/HPC will lead to the displacement of
Source: cryptonews.net
