CleanSpark: The AI Pivot Is Not Worth The Risk

Summary CleanSpark is transitioning from pure bitcoin mining to a hybrid model with significant AI/HPC data center ambitions. CLSK currently earns 100% of revenue from bitcoin mining; AI data center revenue remains uncontracted and future-facing. Valuation already prices in substantial AI/HPC optionality, but execution risks remain until tenant contracts and recurring cash flows are secured. I begin coverage with a hold rating: CLSK's strategic assets are strong, but financials remain volatile and tied to bitcoin cycles. CleanSpark ( CLSK ) is up nearly 50% YTD as the company is making a historic pivot toward being a data center owner and operator. This is despite the fact that the company earns all of its revenue as a bitcoin miner and is highly subject to the ebbs and flows of bitcoin prices. In response to this pivot, the stock has more than doubled in just over two months. Normally, when I see something like this, I begin to think that the market is in over its head, but after breaking down the numbers and running a sum of the parts, I'm thinking the market could actually be somewhat conservative with the current valuation. I begin coverage with a Hold. CleanSpark's Recent Pivot CleanSpark is a domestic digital infrastructure company whose operations are primarily focused on bitcoin mining. This is relatively straightforward as the company deploys ASIC miners across owned and leased data center locations. As time progresses, these miners earn bitcoin rewards like any standard bitcoin miner. However, the company has recently pivoted to categorizing themselves as a data center developer. I want to be very careful with this designation because it is not entirely accurate. This pivot began in October when the company put out a formal press release saying that they had a new Senior VP of AI Data Centers and that they will be transitioning into an infrastructure-first model. This was followed by their November monthly update where they noted that they had acquired 271 acres of land and 285 MW of long-term power agreements near Houston, Texas for a dedicated AI data center. Importantly, the company noted that the energization of more than 200 MW is expected in the first half of 2027. Since that point, there's been two other assets. In January, the company announced it was acquiring 447 acres in Brazoria County, Texas (Houston area), with a power load of 300 MW and the option to expand to 600 MW. In FQ2, the company noted that Sandersville, Georgia is now being developed. There is 250 MW in gross capacity at this facility and CleanSpark is attempting to convert it from a bitcoin mining facility to an AI/HPC data center. A conversion like this, assuming the company could sign a lead customer for it relatively quickly, is likely going to take anywhere from 12-24 months to actually complete. Investors need to understand that if the company converts these facilities, then two things are going to happen. First, there will be outages in the mining rate per day of bitcoin, which will negatively impact the company's financial performance while they attempt to do the conversion. Second, the forthcoming data center revenue will not be incremental, it will be partially offset by the loss in bitcoin mining revenue. The company controls 1.8 GW of contracted power, of which 808 MW is currently being utilized Investor Presentation It's worth noting that the company owns some of the underlying real estate with respect to its bitcoin assets, but it's a very small portion. The company owns the ASIC miners and related infrastructure deployed at current mining centers, but they do not, in most cases, own the underlying land. Coming out of Q1, the company earns no revenue from data centers. They earn 100% of their revenue from bitcoin mining. That means the economics of this business are primarily driven by the amount of bitcoin mined per day, the power cost associated with producing the bitcoin (the largest outright operating cost), the efficiency of the ASIC miners, and uptime/network difficulty. Q2 Likely To Miss But Focus Is on AI Given the heavy bitcoin focus, it's worth recognizing the trading relationship the stock has to bitcoin. It's a correlation just shy of 60% over the past couple of years, but there have been meaningful pockets of deviation, especially in the last few months. As bitcoin has continued to move lower, CLSK stock has actually moved higher and done relatively well. Author Compiled If we take a step back, this is really not the name to go after if investors are looking for an earnings compounder or a sleep well at night name. This is a highly volatile financial profile. For example, in FY25 revenue increased 102% to $766.3 million from $379.0 million in FY24 while adjusted EBITDA printed multiples higher at $823.4 million versus the prior year's $245.8 million. These are great numbers, but mostly a function of bitcoin price appreciation pass through. Fast forward to the current year, things have started to take a turn for the worse. Results in the YTD period are down 8% on revenue and the price of bitcoin is near the lows for the year, which is going to put pressure on financial results. What is helpful from this company is that they do monthly updates based on the price of bitcoin so mapping results is relatively easier to try and come up with estimates. I think it's worth noting, too, that the company holds some bitcoin in a treasury. After the miners receive a bitcoin reward, most of the bitcoin is sold, but a certain amount of bitcoin is put into the treasury in order to benefit from future price appreciation. These are essentially dormant assets with high directional risk on bitcoin. That means during times of bitcoin upcycles, the company will have large mark to market gains and vice versa. The current value of the treasury at the market price is worth just over $860 million. If we take the April and May updates the company has provided for mining revenue, then we can start to formulate a top-line estimate for FQ3. In April, the company produced 640 bitcoin at a price of $74,607/bitcoin. In May, the company produced 671 bitcoin at a price of $79,934/bitcoin. June, so far, has been $63,900 average price. It's been considerably lower as geopolitical events have weighed on prices. Assuming the company could be on par with at least April and May production, I am assuming their 21.5/bitcoin day rate holds and June produces 645 bitcoin. From there, we can calculate the sum product to be $142.7 million in revenue for the quarter. There's still a few days left in the quarter, so this estimate is subject to change, especially as bitcoin remains volatile. This is versus a street revenue estimate of $154.7 million. So, this means about a $12 million miss is coming in FQ3. To be fair, this is a stock that is no stranger to misses - the company rarely hits the mark of street because street fails to react fast enough to update their pricing models. Seeking Alpha The other side of the coin here is that we need to value the 2 GW development platform. This is to ask the question, "What's their AI/HPC data center pivot assets worth net of bitcoin capacity losses?" This is less complicated than what it may seem like. Establishing a revenue/MW/year assumption helps to put some figures in the ground. I am using a range of $1.45-$1.85 million in revenue per MW per year. Using comps for recent contracts, for a 250 MW facility, at an average price of $1.65 million/MW/year, we shake out to about $412 million/year in revenue for the company. Considering the company has about 2.1 GW of power capacity outlined, then scaling up the revenue, this at full operational efficiency is worth about $3.5 billion per year in revenue. Now, we have to have two considerations - 1) how long will it take them to produce this revenue and 2) what's the loss in potential bitcoin mining revenue? Sealy is meant to have the first 200 MW come online in 1H 2027 per the company. Brazoria Phase I at 300 MW is ERCOT approved but it's likely going to be a 2028 story. Brazoria II is likely a 2028/2029 story at the earliest. Sandersville is already built as it is a conversion, there's just no tenant and as I previously outlined, it'll take 14-18 months post-lease. We can make the assumption that they can monetize the rest of their 2.11 GW development platform by the end of the decade. First, the $3.5 billion in data center revenue will likely come at a rather high EBITDA margin. I am assuming 50%. Then, I am applying a base case exit multiple of 10x EBITDA to this. This creates a future value of $17.5 billion and a present value, when discounted over 5 years at 12% (as this is still just an option and very developmental, I am using a higher discount rate), of $9.9 billion. Then, the issue is that we know 808 MW is currently being used out of their 1.8 GW portfolio for bitcoin operations. Using $800k/MW value in lost bitcoin mining revenue, we arrive at a future value of about 10x $646 million which equates to $6.46 billion. Discounted back, this is worth $3.67 billion today. That means the net value of the data center pivot option is worth about $6.26 billion. This assumes they're able to find tenants and contractual rates don't lessen from the current level. I believe that is the largest risk investors assume today. The Value Is Real, But Are The Risks Worth It? In thinking about the balance sheet, the company has long-term debt of $1.79 billion and short-term debt of $2.48 million. Coming out of Q2, the company had $260.3 million of cash, putting net debt at $1.53 billion. With that being said, I think the valuation should not be done on a relative front and should, instead, be done on a sum of the parts basis. First, the bitcoin treasury has a very observable market value considering bitcoin is liquid and trades every minute of every day. The company has about 13,470 bitcoin in their treasury. At the time of writing, the price of bitcoin is $63,850. This implies that the bitcoin holdings are worth $860.1 million. Second, the mining business is worth something. CleanSpark has competitive power costs, decent scale, and is relatively liquid so it should command a decent multiple for its mining operations. The way this is done is by looking at normalized free cash flow from the mining business as it gives us a better representation for the income the business produces ex-bitcoin being saved to the treasury. At the current pricing level, annualized normalized free cash flow is total revenue less power costs, opex, and maintenance capex. The latter is very difficult to estimate as there's not great disclosure around it. Normalized EBITDA at current prices is about $70 million. After subtracting maintenance capex we get to likely $20-$40 million. Giving this a 10x multiple, I am valuing this part of the business at $300 million. It makes very little difference what multiple you use here as the amount is inconsequential relative to the value of both the bitcoin treasury and the AI data center option. Third, the AI data center option certainly needs to be factored in. As I outlined in the prior section, I believe it is worth $6.26 billion today. When we sum this up it equates to the following: AI data center option - $6.26 billion Bitcoin Treasury - $860 million Mining Business - $300 million The total value of CleanSpark - $7.42 billion After subtracting net debt of $1.53 billion, we get an implied equity value of $5.89 billion. The current market cap is just over $4.4 billion, which means that this has upside of 34% to $23/share. Now, there are a few issues with this. First, essentially the entire value of the stock is predicated on this data center option. This company has to make a complete pivot in its operations in order to have the valuation be viable. There is no guarantee they sign tenants and there is no guarantee that they keep costs under control, let alone remain on time, while they try and develop these data centers and conversion projects. It will be years before any investor sees a dollar of revenue from this pivot. So, while the upside figure is enticing, the risks, in my view, are so incredibly high that I don't think this is a worthwhile endeavor. Layer on top the fact that I think the company is missing Q2 earnings based on prevailing bitcoin prices. Also, I believe that the stock is still likely going to be a short-term bitcoin proxy, like it has in the past, so investors are subject to just a crypto pass through while they wait for the data center optionality to come to fruition. Passage Research Risks As I am beginning coverage with a Hold, risk can cut both ways. The biggest risk to the upside is that the company is able to monetize its data center revenue assets faster and or obtain more assets in this bucket in the near-term. These assets come at a net benefit to the enterprise in terms of incremental revenue and cash flow. The more the business transitions towards this type of a business mix, the better. This is not to say that in the short-term, announcements around things like who the company ends up signing in terms of a tenant can't move the stock. At the same time, there is the risk that contracted power stops commanding a pricing premium as AI exits an early adoption stage. Additionally, the higher this ends up trending over time as a percentage of the total business' cash flow, the generally less volatile shares should become. Being the lessor of a data center means that they'll end up adopting a more favorable structure and provide shareholders with greater cash flow visibility. Today, and for the foreseeable future, the company is going to still be getting the very vast majority of its revenue from bitcoin mining. This means that the company will still be subject to the ebbs and flows of the price of bitcoin. Conclusion CleanSpark is certainly an interesting bitcoin miner as they've now decided to make a pivot towards AI data centers. In doing this, they're targeting a much more stable, yet fast-growing part of the market that can help their company's financial standing for years to come. The issue the market is going to have from here is how to appropriately value the data center option. In making my attempt at it, I do see a path for upside, but I believe the risks are not worth the reward and am exercising caution at this time.