It has been nearly two years since I last covered Core Scientific (NASDAQ:CORZ) for Seeking Alpha. In the time since that “sell” call, CORZ shareholders have been on quite the journey to say the least.
Shortly following my late-October 2022 CORZ article, the crypto market experienced the collapse of FTX which further depressed prices of crypto coins and proxies like Bitcoin (BTC-USD) mining stocks. For Core Scientific specifically, the company filed for Chapter 11 bankruptcy and re-emerged with a reorganization plan highlighting reduced debt and 60% equity retention for shareholders of the delisted stock. It admittedly could have gone far worse for CORZ holders and the relisted equity has performed quite well this year.
More recently, shares have CORZ have rocketed higher largely in response to the company’s efforts in the AI datacenter space and a buyout offer from its primary HPC services client. In this update, we’ll look at the CoreWeave news, financials as of Q1-24, and the company’s mining business.
Core Scientific and CoreWeave
On June 4th, Bloomberg reported an offer from CoreWeave to buyout Core Scientific at a roughly $1 billion valuation, or $5.75 per share. This came just one day after Core Scientific announced contracts for a large expansion of the company’s HPC services agreement with CoreWeave. CORZ shares have rallied nearly 50% on this news as of article submission and are actually trading at almost about a 20% premium to CoreWeave’s offer price. Per the company’s June 3rd press release, the newly announced HPC contracts are for 200 MW of HPC infrastructure and will last 12 years with the potential to expand further. According to Core Scientific, these agreements will result in approximately $290 million of recurring, dollar-denominated revenue annually.
This is an expansion of a pre-existing relationship between Core Scientific and CoreWeave. Between 2019 and 2022, CoreWeave was utilizing Core Scientific for GPU hosting. More recently, Core Scientific deployed 16 MW of capacity at its Austin, Texas data center for CoreWeave earlier this year. It’s important to point out that GPUs and ASICs aren’t interchangeable. To scale HPC service revenue, Core Scientific has to modify some of it’s existing infrastructure to better align its mining and HPC segments. Core Scientific is estimating that $300 million of the capital investment required for infrastructure modifications will be funded by CoreWeave. Site modifications are expected to begin in the second half of this year with operations beginning in the first half of 2025.
Q1-24 Earnings & Balance Sheet
The news of both a possible buyout from and business expansion plans with CoreWeave follows what has been a fairly impressive turnaround for Core Scientific in recent quarters. At the end of March 2024, the company reported over $173 million in revenue against $101.6 million in costs of revenue and $17 million in total opex:
Factoring in over $111 million in net reorganization items that include extinguishment of loans, leases, and notes, net income in the quarter came in just under $211 million:
This was the company’s first positive quarter for net income since December 2021. Breaking out the revenue by segment shows Core Scientific is still largely dependent on digital asset mining as it accounts for nearly 84% of the company’s total revenue with hosting making up the rest:
10-Q (Core Scientific )
In Q1-23, Core Scientific had about a 14% gross margin on hosting revenue from customers. That figure more than doubled to over 31% in Q1-24. On the surface, we see similar gains in gross margin from the mining segment with 26% gross margins last year improving to roughly 45% in the last quarter. We’ll explore the mining segment a bit more in the next section.
Briefly shifting to the balance sheet, Core has about $98 million in cash with $814 million in total assets against more than $1.1 billion in total liabilities.
May Investor Deck, Slide 21 (Core Scientific)
The good thing is most of the company’s debt isn’t due for several years with more than a half billion dollars in principal weighted to 2027 or after. Given the company’s apparent HPC revenue from CoreWeave ramping up in 2025, Core Scientific’s financials aren’t actually as bad as I previously suspected. However, Core Scientific’s future is seemingly at the mercy of CoreWeave; which from where I sit explains why CoreWeave’s buyout offer has come in shortly following the expanded HPC services relationship. There is presumably far more upside for CoreWeave if it simply owns the infrastructure rather than renting it from Core Scientific.
Bitcoin Mining Business
Frankly, the HPC services agreement with CoreWeave probably couldn’t be coming at a better time given the recent halving of the Bitcoin block reward subsidy. We don’t yet have May production data from the company, but the trend over the last several quarters is pretty clear. As global Bitcoin hash rate grows, share of the block reward declines if miners can’t continue to scale owned production capacity in line with global hash:
BTC Production (Author’s Chart, Core Scientific Filings)
Since January 2023, Core Scientific’s monthly BTC production has been in decline even as self-mined EH/s has continued to grow.
BTC per EH/s April (Author’s Chart, Company Filings)
In April, Core Scientific’s Bitcoin produced per EH/s was in the middle of the pack compared to most of the top public Bitcoin mining companies. If we adjust the April data down by 50% to provide an estimate for May and assume revenue from transaction fees continues to be sporadic, we get 19.7 BTC per EH/s going forward. Using Core Scientific’s April mining capacity, we can estimate 402 BTC in May. Short of higher BTC prices in perpetuity or meaningful growth in transaction fees, Core Scientific’s Bitcoin mining business is likely to deteriorate in the months and quarters ahead. All this said, the company’s mining segment may indeed benefit from higher BTC prices later this year and next year.
Closing Summary
If we take my May BTC production estimate of 402 for Core Scientific and build an annualized revenue table based on Bitcoin’s price, it takes a six figure BTC price to get close to $500 million in full year revenue from mining:
| BTC Price | Revenue per month | Annualized |
|---|---|---|
| $70,000 | $28,140,000 | $337,680,000 |
| $80,000 | $32,160,000 | $385,920,000 |
| $90,000 | $36,180,000 | $434,160,000 |
| $100,000 | $40,200,000 | $482,400,000 |
Source: Author’s calculations, assumes 402 BTC mined per month
Making the assumption that BTC’s price will hit $100k by 2025 and assuming $290 million in HPC service revenue from CoreWeave in the year as well, Core Scientific’s stock trades at about 1.5 time sales at $7 per share. This would be roughly a 47% discount to info tech sector median. Given that, I’m actually not all that surprised to see the stock trading at a premium to CoreWeave’s buyout offer.
I’ll close with saying I don’t have a dog in this fight. I’ve never shorted CORZ despite calling it a “sell” two years ago. Despite the buyout offer, I do think there is still risk in this name. I’m personally of the view that “AI” as an investment idea has skewed more to hype than real substance. Given that, this rally may be an incredible opportunity for shareholders from the previous cycle to get out alive from what was a bankruptcy story not too long ago.
Read the full article here
