Core Scientific: HPC Pivot Begins To Bear Fruit

Summary Core Scientific’s stock has nearly doubled since June, outperforming the market and tech giants like NVIDIA. Despite a significant debt reduction and improved balance sheet, Core Scientific’s growth heavily relies on its relationship with CoreWeave, raising concerns about client diversification. Bitcoin mining remains Core Scientific’s primary revenue source, but declining BTC production and high operational costs highlight the need for growth in the HPC segment. The delay in CoreWeave’s 118 MW option and the lack of additional HPC clients make me cautious about investing in Core Scientific at this time. It has been slightly over 4 months since I last covered Core Scientific ( CORZ ) for Seeking Alpha. At that time, my rating of the stock as a ‘hold’ was mainly derived from the company’s high debt position, as well as my feeling that “AI” revenue broadly would fail to live up to the lofty hype. Smarter than any one man, the market disagreed. Here’s the performance of CORZ since my June 4th article was published: Data by YCharts Not only has CORZ nearly doubled since my last piece, but it has essentially outperformed everything . It has outperformed the broad market, Bitcoin ( BTC-USD ), mining peers , NVIDIA ( NVDA ), and even MicroStrategy ( MSTR ) since early June. So, let’s revisit. In this update, we’ll get into the company’s Q2 earnings, positive signs pertaining to the balance sheet, and look at what could work against Core Scientific’s rapid revenue trajectory. Q2 Performance For the three months ended June 30th, Core Scientific reported $141 million in total revenue. Notably, this was the first quarter the company generated revenue from its HPC hosting segment. That number came in at $5.5 million in the quarter and makes up a small portion of the overall revenue story to this point: Second Quarter 2024 10-Q (Core Scientific) At $110.7 million in the quarter, the company is still highly reliant on Bitcoin mining for top line revenue growth. The 11% year-over-year growth in total revenue was mainly attributable to the 14% growth in Core’s self-mining business. Hosted revenue actually declined by roughly 6%. Gross margin in the digital asset mining business came in at 28% which was a slight reduction from 31% in Q2-23. Margin on the HPC business was the smallest of Core Scientific’s three segments, at just 11%. Quarter ended June 2024 2023 YoY Gross profit $38,817 $36,959 5.0% Total Opex $31,383 $27,120 15.7% Operating income $6,579 $9,469 -30.5% Source: Core Scientific, 10-Q, dollars in millions After opex, Core’s operating income for the quarter comes in at a little under $6.6 million – down more than 30% from the prior year. A large part of this decline in operating income is the faster growth in Core’s cost of mining relative to its revenue from mining. I suspect this trend will only continue given the dynamics from the halving event in April: Monthly Bitcoin Production (Core Scientific, Author’s chart) BTC production on a monthly coin basis is down quite a bit from last year. In Q2-23, Core Scientific mined 3,470 BTC. That figure was down to 1,680 BTC in Q2-24 and will be even lower when the company reports Q3 earnings. Still, Core Scientific’s pivot to HPC services should help alleviate the potential stagnation in Bitcoin mining revenue if the coin price fails to meaningfully appreciate from the current $66k level between now and the end of the year. Q2 Earnings Deck, Slide 15 (Core Scientific ) Core Scientific is guiding for 382 MW of dedicated HPC infrastructure to CoreWeave through the first half of 2026. As I’ve mentioned in other articles for Seeking Alpha, the HPC data center business is something that almost every publicly traded Bitcoin miner is now actively pursuing, with only a handful generating revenue from the segment. Core is now one of those companies and seemingly has one of the biggest whale buyers of compute in CoreWeave. Debt Repayment At the end of March, Core Scientific had over $600 million in total debt, with just $98 million in cash. In my view, the company is in a much better financial position today after a series of moves. Q2 Earnings Deck, Slide 10 (Core Scientific) In early-July, Core Scientific had a $260 million debt reduction via mandatory conversion of debt notes into 45 million common shares. Subsequent to the July debt conversion, Core Scientific issued $400 million in new convertible notes at 3% set to mature on September 1st, 2029. The proceeds of these new notes were used to pay down $267 million in debt at a higher rate of interest, which Core Scientific says was approximately 12%. The improvement in the balance sheet has come at a slight dilution cost to equity holders: Data by YCharts As of August 2nd, 2024, there were 285.2 million shares outstanding following the conversion to common. 40 million shares were added due to the mandatory conversion, and an additional 35 million common shares have been added from exercised warrants. This is roughly a 40% increase in common shares outstanding in a single quarter . The net loss in Q2 was roughly $800 million, but almost entirely attributable to a $796 million non-cash adjustment to Core Scientific’s warrants because of the increase in value of the company’s equity. CoreWeave Relationship With respect to Bitcoin mining, the major selling point for Core Scientific investors at this point has to be the company’s HPC prospects. This segment can potentially generate $6.7 billion in total revenue over the next 12 years, all from a single client. That client is CoreWeave. Recall, CoreWeave attempted to buy Core Scientific outright at a $1 billion valuation earlier this year. CORZ is now worth over $3 billion. Q2 Earnings Deck, Slide 14 (Core Scientific) CoreWeave’s interest in the company probably shouldn’t come as much of a surprise, given Core Scientific’s 80% guided margins on the current contract between the two entities. With CoreWeave funding capex and absorbing the cost of power and utilities, it’s no wonder that the CoreWeave wants to buyout Core Scientific. Perhaps a compelling development then is the extension of CoreWeave’s option for 118 MW . On the conference call in August, Core Scientific CEO Adam Sullivan made note of the option expiration in early September: We are in discussion with CoreWeave and a number of other potential clients to contract the remaining 118 megawatts of our infrastructure available to support additional HPC hosting. Our clients’ option for additional infrastructure extends through the beginning of September, providing the ability to contract additional capacity at the original contract terms. The two companies agreed to a 30-day extension of that 118 MW option, yet as of October 15th, there has been no update to my knowledge. I don’t want to speculate as to why this option wouldn’t be exercised if it were indeed to fall through. However, I don’t think it would necessarily be a good thing for CORZ unless it leads to a new client for Core Scientific. On October 11th, CoreWeave announced a $650 million credit facility to continue AI growth. This puts the total capital raised by CoreWeave at $12.7 billion. Given the capital raise, it’s strange to me that we haven’t yet seen the 118 MW option move forward, even though we appear to be passed the deadline. Closing Remarks Core Scientific’s turnaround remains one of the more impressive survival stories to come out of the 2022 crypto winter. I will fully admit that my caution on CORZ this summer was a missed opportunity. The performance in CORZ this year and over these last few months speaks for itself. All this said, I’m still not willing to jump into this stock for one simple reason; growth appears to be very dependent on CoreWeave today. Though Bitcoin mining still makes up most of Core Scientific’s revenue, mining is a deeply flawed business model without exponential growth in coin prices or transaction fees. The latter of which is unlikely given the biggest supporters of the asset ‘HODL’ it rather than transact with it. Thus, the necessity to grow HPC. Given the expertise in data centers, this is theoretically a complimentary business. What concerns me about Core Scientific’s HPC growth is the lack of client diversification. To their credit, management seems well aware of this Achilles heel and is actively addressing it. Given that, I think it’s interesting that CoreWeave’s 118 MW option is taking as long as it has been, especially since the company just completed a capital raise. In my view, Core Scientific is definitely one for a watch list. But I can’t say that I would get long the stock until I knew there was another HPC client.

Source: Seeking Alpha