Summary CleanSpark is pivoting from pure-play Bitcoin mining to data centers, aiming to capture AI/hyperscaler demand and improve cash flow sustainability. Despite record FY2025 revenue of $766M and positive operating income of $319M, CLSK's free cash flow remained deeply negative at -$461M, driving continued dilution and increased debt. Management turnover and a $1.15B convertible bond raise signal strategic urgency, with initial data center conversions targeting Sandersville and Texas facilities (200MW each). Full conversion of 1.4GW capacity could yield $1.4B–$2.1B annual revenue, but immediate, complete transition is not the current outlook. CleanSpark ( CLSK ) is a Bitcoin ( BTC-USD ) miner in transition to using its power capacity for AI data centers. It's a late shift and change of tone from previous guidance, and it's one worth a follow-up. While I think this will be a better business over time than BTC mining, tight cash flows and capital needs make the current valuation unfavorable. Summary of Previous Thesis When I wrote about CleanSpark in March, I was cautious about its status as a BTC mining pure play. This was something that made the company stand out, as other BTC miners were converting many of their assets into data centers. The move, broadly accepted across the BTC mining industry, leverages electrical capacity for an end use that can produce reliable free cash flow. CleanSpark's decision to stay out of it was, therefore, a curious one. Screenshot from previous thesis (Author's display of 10K data) Seen above, their BTC mining business was a cash burner, regularly requiring more shares to be issued, in order to finance operations. CleanSpark was betting that competitors' shift to data centers would make their mining operations more competitive, as it would give them a greater share of the global hash rate and therefore improve their BTC output. Nevertheless, BTC still needed to appreciate considerably to turn cash flows positive. Lacking those in the meantime and threatened by continual dilution that would kill off realized upside, I gave CLSK a cautious Hold rating. Updates and Subsequent Events In the months that followed, CleanSpark proceeded with the BTC mining pure play approach. They expressed growing confidence that the business would become self-sufficient, announcing a financing deal with Coinbase ( COIN ) in April. They stated that a portion of BTC mined would be used to sustain operations. Later, they touted reaching 50 exahashes per second in their mining. CEO Zach Bradford stated: As we look to the future, CleanSpark has laid the foundation to achieve 60 EH/s and beyond. This expansion capacity aligns with the Company's broader strategy to grow with capital efficiency only possible for an operator that has achieved escape velocity . "Escape velocity" was a term used in both cases, expressing that CleanSpark was near to achieving what I was afraid they would not achieve: an end to cash burn and shareholder dilution. Cash Flow Statements (Q2 & Q3 Fiscal 2025 Forms 10Q) In spite of this, there was still a total cash burn of $234M in fiscal Q3, after accounting for operating cash flow, purchase/disposition of miners, BTC proceeds, and purchase of fixed assets. If Bradford believed CleanSpark was reaching escape velocity, the numbers were not yet bearing it out. Whether he was right or wrong, Bradford resigned as CEO shortly after fiscal Q3 results were released. He was replaced by Matt Schultz, Bradford's fellow co-founder and predecessor as CEO (from 2014 to 2019). In early September, new management announced a strategic pivot to data centers. CleanSpark was finally on the AI-hyperscaler train. Stepping outside of self-funding operations, CleanSpark began raising capital again, such as a recent $1.15B convertible bond offering in November . End of Fiscal 2025 Financials Fiscal 2025 was, in large part, a record year for the company, with revenue reaching an all-time high of $766M. This also supported operating income finally turning positive, reaching $319M. Income Statement (Fiscal 2025 Form 10K) For many folks this is a welcome sign, but the cash flow situation reveals some underlying cracks, coming out of the year. Cash Flow Statement (Fiscal 2025 Form 10K) While cash burn was lessened somewhat by reduced capex on miners and liquidation of BTC, operating cash flow still decreased overall for the year, reaching -$461M. BTC 1Y Price History (Seeking Alpha) Some likely recall the recent decline of the price of BTC in the fall. For CleanSpark, the fiscal year ended before this drop. They were mining BTC at prices above $100K for a large portion of the year. Fiscal 2025 Form 10K In order to raise capital and plug the gap, CleanSpark shares outstanding, basic and diluted, have both grown in the year, so the consequences of dilution persisted. Liabilities (Fiscal 2025 Form 10K) They ended the year with significantly more debt, largely from the credit facilities they started throughout the year. Then there are the convertible bonds issued later that I mentioned earlier. Overall, it's a somewhat improved picture but still a company with serious capital needs. Outlook and Valuation CLSK trades for about $10 per share and a market cap of $2.6B. As free cash flow has been negative, the task is to determine how following the shift to data centers can tip the scales into positive cash flow. Much of what I will say next will work with very general estimates, given the lack of specifics at this stage. To start, I've looked at similar deals by other companies to give us an idea of what rates they could make with their electrical capacity: Applied Digital's ( APLD ) contract with a hyperscaler Another deal between APLD and CoreWeave ( CRWV ) Ionic Digital's contract with Nscale These contracts show that a data center can make about $1M to $1.5M per MW each year. I've included a table for your convenience. Author's display of data center revenue estimates Total electrical capacity was reported to be 1.4 GW in November. If we assume their total portfolio of capacity were used for data centers today, one could imagine revenue of $1.4B to $2.1B annually, dependent on the rates of local markets and the like. Of course, that assumes complete and immediate conversion of these assets, which isn't the outlook. During the Q4 earnings call , management largely highlighted the Sandersville and Texas facilities as the first places that could be converted to AI/HPC uses. Q4 Fiscal 2025 Investor Presentation These two sites are reported to have about 200 MW of capacity each that could be active by 2027. They could contribute about $400M to $600M in revenue. As the business currently has gross margins of about 55% just on its BTC mining, the underlying assumption is that the average margin for AI/HPC would be meaningfully superior to this. Income Statement (Fiscal 2025 Form 10K) The devil is going to be in the details, but a look at key expense items currently show that these assets could produce an additional $100M to $200M in operating income. The upfront capex, however, means that free cash flow would likely not turn positive for some time, however. Liquid Assets (Fiscal 2025 Form 10K) As CleanSpark ended fiscal 2025 with only $43M in cash and is slow to liquidate its $967M BTC holdings, this is why they had to do the $1.15B convertible bond offering in November and have leaned into their BTC-backed credit facilities instead. The cash from the convertibles should be sufficient to cover capex needs for the next couple of years (if recent years are a guide). Additional cash burn from operating expenses will force them to raise more capital at some point, and this will likely entail some kind of a dilutive offering. When I see a market cap of $2.6B, I think the fundamentals should practically produce a cash flow multiple of about 10 within a few years. The timing of things makes it seem unlikely they will hit such a milestone in a few years without diluting more. The guidance in the call, after all, was: Our focus is on building a capital stack, which minimizes dilution . Management realizes what I realize and can only promise to dilute as little as possible. I think CLSK needs to be more conservatively valued for it to reflect a real shot at undervaluation. A market cap of $1.5B, which would make an entry price of $5.80 or less, is one that seems to allow for a slower realization of positive cash flow and likely some kind of margin of safety. That isn't to say that a very patient investor couldn't see upside on CLSK at the current valuation, but we're likely thinking very far into the future and imagining that execution goes very well for the most part. Fiscal Q1 results, as well as monthly updates of power capacity, will be important in assessing just how large the pipeline is and how much higher-margin revenue could be realized that isn't on the table right now. Conclusion CleanSpark likely made a pivot that needed to be made, even before BTC prices declined. Nevertheless, strategic transitions take time and cost money. CLSK meanwhile is valued in the billions, when generating cash in the hundreds of millions without more dilution remains an open question. Investors can simply wait for a more reasonably priced entry, and for that reason, I maintain my previous Hold rating.
