Our CLSK stock analysis starts with a contrast nobody on the company’s official communications has fully acknowledged. Through fiscal year 2025 — which for CleanSpark ended September 30, 2025 — management talked about disciplined, US-only Bitcoin mining as a thesis. Then, almost without breaking stride, the company signed two $100 million HPC credit facilities in September 2025, bought 271 acres in Austin County, Texas for an AI and cloud data center campus, bought another 447 acres in Brazoria County, Texas with up to 600 megawatts of power earmarked for AI workloads, and closed a $1.15 billion zero-coupon convertible note to fund the pivot. CleanSpark didn’t ride the bitcoin-miners-becoming-AI-infrastructure wave from the front. It joined the line eight months in.

CLSK

--

Published$14.30·May 19·

1D5D1W1M3M6M1Y

Data: Yahoo Finance

The stock — ticker CLSK — last printed $14.30 going into publication on May 19, 2026, and is up roughly 37% in the trailing twelve months. That number matters because of what it sits between. Pure-play miner Marathon Digital (MARA) is down 25% over the same window. The AI-pivoted miners — IREN, Cipher Digital (CIFR, rebranded from Cipher Mining in February 2026), and Core Scientific (CORZ) — are up 491%, 402%, and 117% respectively. CLSK is the in-between trade: pivoted, but late.

The Setup: What Changed in Six Months

CleanSpark is one of the largest US-focused publicly-listed Bitcoin miners by hash rate, alongside Marathon and Riot Platforms. The company operates twelve mining sites in Georgia (about 300 megawatts of contracted power and counting), three sites in Mississippi (Vicksburg, Meridian, Wiggins), facilities in Tennessee, and a Wyoming campus in Cheyenne that is being commissioned as a fully immersion-cooled site. Operational hash rate reached 50.0 exahashes per second in April 2026 — the highest the company has ever produced — though installed capacity is closer to 27.6 EH/s per the most recent Q2 FY2026 quarterly results.

That backdrop matters because, throughout the fiscal year that ended September 30, 2025, this profile was the entire pitch. CleanSpark was the one major public miner that had not announced an AI or HPC partnership and explicitly differentiated itself on operational focus. By the time the FY2025 10-K was filed on November 25, 2025, that positioning was already moving. The company had executed two $100 million HPC credit facilities in September 2025 and was actively scoping Texas land for AI-ready data center build-outs. By the first calendar quarter of 2026 it had acquired two large Texas sites totaling 718 acres and closed a $1.15 billion 0% convertible note to fund the expansion.

The contrast in messaging is what gives this post its angle. The market spent 2024 and most of 2025 paying up for any Bitcoin miner with a credible AI story — CORZ, CIFR, IREN — and treating CLSK as the disciplined holdout. CleanSpark didn’t volunteer that it had quietly changed its mind until the capital commitments were already in place. The question for retail investors trying to size a position now is whether being late is a structural disadvantage or whether the late entry preserves optionality at lower cost.

CLSK Stock Analysis: What Q2 FY2026 Actually Showed

The most recent quarter — Q2 fiscal year 2026, period ending March 31, 2026 — was not a clean print. Revenue came in at $136 million. Net loss was $378 million, dominated by a non-cash mark-down on the company’s Bitcoin holdings as BTC pulled back during the quarter, plus higher interest expense as the balance sheet absorbed new debt. Total cash and bitcoin on the balance sheet at quarter-end stood at $260.3 million in cash and $925.2 million in HODL value (current bitcoin plus non-current bitcoin plus bitcoin pledged to counterparty collateral arrangements). The Q2 FY2026 results press release details the full breakdown; the April 2026 monthly update reported 13,453 BTC held.

On the operations side, the Q2 print was more mixed than the loss line suggests. Operational hash rate crossed 50 EH/s in April 2026. Total contracted power capacity stood at approximately 853 megawatts across the US data center footprint, with 1.8 GW under contract when including planned AI capacity. The company is currently utilizing roughly 808 MW.

All-in cost per Bitcoin mined is the single most important operating metric for a miner one year past the April 19, 2024 halving (which cut block rewards from 6.25 to 3.125 BTC). The most recently disclosed figure — from Q4 fiscal 2025 — was $52,482 per coin all-in, broken into $36,139 of direct cash production cost and $16,343 of corporate overhead. That figure represented a roughly 25% improvement on the prior period. The Q2 FY2026 10-Q on SEC EDGAR reports $81.7 million of cost of revenues exclusive of depreciation and amortization against $136.4 million of mining revenue, but does not break out the per-BTC figure at the same level of detail as the Q4 FY2025 release.

Here is how CleanSpark sits inside its peer set as of mid-May 2026:

Two things stand out. First, the market has already paid up generously for any miner with a signed hyperscaler contract — IREN, CIFR, and CORZ are the three with concrete revenue agreements and the three biggest re-rates. Second, MARA’s negative one-year return is the cleanest available read on what the market thinks of a public miner that has not made the AI move. CLSK’s position — pivot announced, capital committed, contracts pending — gets the market a portion of the upside and a portion of the downside.

The Treasury Sitting on the Balance Sheet

One reason CLSK’s late pivot is more credible than it would otherwise be is the balance sheet that funds it. As of March 31, 2026, the company reported approximately 13,453 BTC of Bitcoin holdings and $260.3 million in cash. Total HODL value was $925.2 million when including bitcoin pledged to counterparty collateral arrangements. That stack is the implicit reason the $1.15 billion 0% convertible cleared at zero coupon — the lenders priced the conversion-option upside against a balance sheet that already carries hundreds of millions in market-priced digital assets.

For retail investors trying to value the stock, the practical exercise is a sum-of-parts approach. If you mark Bitcoin to the spot price, divide by shares outstanding, and back the result out of the market cap, the residual is what you are paying for the operating miner plus the embedded AI optionality. Different BTC prices give very different residuals — and that elasticity is the point. CLSK’s equity is more BTC-levered than a pure operating-miner valuation would suggest, because the treasury is large relative to the operating cash generation.

The dilution overhang from the 0% convertible is real even at zero interest. The 0% coupon is the price the lenders accepted in exchange for the conversion option — meaning they expect the equity to be worth materially more at exit than the strike, and the share count to grow accordingly. The Q2 FY2026 10-Q is the right place to read the conversion mechanics, lockup terms, and any associated capped-call hedges. We covered similar dilution-overhang dynamics in our ONDS catalyst analysis earlier this week.

The AI Pivot — Late, But With Receipts

Industry context first. The AI infrastructure trade that propelled the miner re-rates of 2025 is built on power. Hyperscalers — Microsoft, Amazon, Google — need gigawatts of contracted, immediately-buildable electricity at sites with grid interconnect already approved. Bitcoin miners have spent five years acquiring exactly that. Converting a miner site to a high-density AI colocation facility is non-trivial (different cooling, different network topology, much higher power density per rack), but the underlying scarce resource — the megawatts and the substation — is already owned.

The peers have monetized that scarcity at scale. Core Scientific has roughly $10 billion in contracted colocation revenue against an aggressive 4.5 gigawatt expansion pipeline; Q1 FY2026 colocation revenue was up over 800% year-on-year. Cipher Digital — formerly Cipher Mining — rebranded in February 2026 and disclosed 600 megawatts of contracted HPC capacity across a 15-year, 300 MW lease with AWS and a 10-year, 300 MW lease with Fluidstack and Google, for $9.3 billion of total long-term AI contracts. IREN booked a $9.7 billion five-year AI cloud agreement with Microsoft in late 2025 and energized its 1.4 GW Sweetwater 1 Texas data center on May 1, 2026.

The relevant comparison for CLSK is not that it has been slower to acquire the optionality — it is that it has been slower to monetize it. We outlined the conversion mechanics in detail in our BITF/KEEL infrastructure thesis from May 14, where Bitfarms spun out the HPC business into Keel Infrastructure and crystallized the value of its power footprint as a separate equity story. CLSK has chosen not to spin out, not to rebrand, and so far not to sign a hyperscaler lease. The Texas acquisitions are real. The contracts to fill them are not yet announced.

CleanSpark’s stated pivot timeline runs as follows:

• September 2025: Two $100 million HPC-focused credit facilities established. This is the first public capital commitment to the AI direction, and it was understated in the FY2025 communications.
• Late 2025: Acquisition of 271 acres in Austin County, Texas, with long-term power supply agreements for up to 285 megawatts to support AI, cloud, and enterprise workloads.
• Q1 calendar 2026: Announced acquisition of 447 acres in Brazoria County, Texas, with transmission-level power enabling 300 megawatts of initial demand load and potential capacity for further expansion to 600 megawatts. The Brazoria County deal was detailed by Data Center Dynamics.
• November 2025: Closed $1.15 billion 0% convertible note. Proceeds earmarked for power and land portfolio expansion.

The honest read is that the pivot has receipts — credit facilities, land, capital — and is missing the one thing that drove peer re-rates: a signed hyperscaler lease that translates contracted megawatts into contracted dollars on a multi-year basis. The CORZ/CoreWeave deal, the CIFR/AWS lease, the IREN/Microsoft contract were each the catalyst that shifted the equity from “Bitcoin miner with optionality” to “AI infrastructure operator with a Bitcoin miner attached.” CLSK has not yet had its catalyst day.

The counter-argument matters. Late entrants into infrastructure cycles often pay premium prices for the worst remaining sites because the best power, the best interconnects, and the best site geographies have already been allocated. Texas in particular has seen a step-change in power pricing for AI-ready sites between 2024 and 2026, and the developer queue at ERCOT has lengthened. The bear case on CLSK’s late pivot is not that it is fake or unfunded — it is that the marginal site CleanSpark is acquiring in 2026 may cost more, deliver slower, and contract at thinner margins than the marginal site CORZ acquired in 2023 or IREN in 2024. Whether that is true depends on the specific land, power, and interconnect terms of the Brazoria and Austin acquisitions, which are not fully disclosed.

Forward Catalysts We’re Watching

The next twelve months will tell us whether CLSK’s pivot crystallizes into contracted revenue or stalls into another year of capex commentary. The catalysts in order of materiality:

• A signed hyperscaler lease at either Texas site. This is the catalyst that re-rated CORZ, CIFR, and IREN. Until CLSK signs one, the equity trades on optionality, not on cash-flow visibility. Investors should verify CleanSpark IR for any announcement.
• Q3 FY2026 print (calendar Q2 2026, expected report in August). Two specific numbers matter: all-in cost per Bitcoin mined (did the 25% improvement trajectory from Q4 FY2025 continue?), and any disclosed capex pacing on the Texas sites.
• First energization milestone at Austin County or Brazoria County. Operating power at an AI-ready site is what converts contracted megawatts into the kind of capacity hyperscalers will sign against.
• Bitcoin price action. CLSK’s $925 million HODL stack is mark-to-market, so the equity carries a direct beta to BTC spot. A 20% BTC drawdown hits reported earnings; a 30% rally adds hundreds of millions to book value without any operating change.
• Convertible dilution timing. The Q2 FY2026 10-Q is the right place to read conversion price, lockup, and any capped-call structure on the $1.15 billion 0% note.
• The 2028 halving cycle. Roughly two years out, but mining economics already start to price the next reward cut about a year before it happens. CLSK’s relative position against a halved block reward will depend on how much of the Texas AI revenue is online by then.

Bottom Line

CLSK is the in-between trade in the Bitcoin miner cohort. The pivot is real, funded, and now visible in the capital base — but the contracts that re-rated CORZ, CIFR, and IREN have not yet shown up at CleanSpark’s door. Retail investors who want Bitcoin mining beta with a sizeable HODL treasury and a real-but-late AI optionality kicker may find the in-between positioning attractive precisely because the re-rating is still ahead. Retail investors who are buying the cohort to ride the AI infrastructure trade specifically should note that the trade is already in CORZ, CIFR, and IREN, and CLSK is starting from behind.

The single thing we are watching in the next quarterly cycle is whether CleanSpark announces a hyperscaler lease at either Texas site. That announcement — or its absence — is the data point that resolves the late-mover question.