CleanSpark: Sandersville Is Real. Texas Is the Risk.

The market is still pricing CleanSpark like a bleeding Bitcoin miner. B. Riley's $26 target says something different.

The disconnect is wider than the target itself. The real question is whether the Sandersville lease — the one signed, contracted piece of this puzzle — justifies nearly doubling the stock from $13.50. The Texas campuses, which carry the bulk of the valuation upside, are letters of intent, not revenue. And as of this week, Texas regulators have frozen new data-center grid connections in a moratorium that could delay them further.

The setup is worth looking at because the Sandersville deal is real and the market hasn't priced it in yet. But the path from $13.50 to $26 requires Texas to cooperate, construction to stay on schedule, and the mining bridge to hold until data-center cash flow arrives.

The old story versus the new one

CleanSpark lost $239.8 million in the third fiscal quarter, which ended June 30. Revenue fell 30.5% year over year to $138 million. The stock trades below $14, down more than half from its 2025 highs. Short sellers have loaded up — 33% of the float, the highest among crypto companies above $2 billion in market cap.

The headlines are about mining economics getting worse. Gross margin compressed to 38% from 40%, squeezed by rising power costs. The quarter-to-quarter improvement in the net loss was driven by a smaller mark-to-market charge on Bitcoin holdings, not operational gains.

But management reported core adjusted EBITDA... at positive $20 million. Digital asset management generated $25.8 million in net cash year to date. The mining operation, while unglamorous, hasn't collapsed. It's holding as a bridge.

The bridge is meant to carry the company through to when the Sandersville data center in Georgia begins generating revenue in the fourth quarter of 2027. That facility, the first of its kind for CleanSpark, is covered by a 20-year triple-net lease (the tenant covers taxes, insurance, and maintenance) with an unnamed high-investment-grade technology company. The deal locks in approximately $6.6 billion in contracted revenue, with extension options that could push that toward $11.6 billion. Average annual net operating income is projected at roughly $330 million.

That single deal changes the risk profile. A triple-net lease at that scale, with an investment-grade counterparty, is durable infrastructure cash flow. CleanSpark has fully funded the equity portion of construction and ordered all long-lead items. They're targeting project-level debt for more than 90% of remaining costs to avoid diluting shareholders.

B. Riley's math

B. Riley's sum-of-the-parts model breaks the $26 target into pieces. Sandersville is valued at $6.27 billion using a 20x multiple on modeled annual NOI of $314 million, with $1.93 billion in capex deducted. The remaining $10.91 billion comes from the Texas portfolio: Sealy at $3.89 billion and Brazoria at $7.02 billion, both valued at 14x on modeled NOI. After deducting $1.53 billion in net debt and adding back $905 million in Bitcoin holdings (13,924 BTC at $65,000 per coin), the model yields $25.76 per share on 257 million fully diluted shares.

The multiple gap tells you where B. Riley sees the optionality. Sandersville — the contracted asset — gets 20x. The Texas sites, which are covered only by letters of intent and exclusivity arrangements, get 14x but represent the largest share of the total equity value. In other words, more than half the target depends on Texas deals that haven't been signed.

The Texas risk

This is the hinge. CleanSpark's Sealy campus (285 MW) and Brazoria campus (up to 600 MW in two phases) sit near Houston. Both are ERCOT-connected, meaning they need grid interconnection approval from Texas regulators.

On August 3rd — the same day B. Riley published its report — Governor Greg Abbott announced a moratorium freezing approval of new data-center grid connections across ERCOT. The freeze requires comprehensive audits of every project in the interconnection queue, covering tax incentives, water usage, power demand, and community impact. No timeline has been set for when the audit concludes.

CleanSpark's earnings call on August 6th acknowledged the regulatory headwind. Management noted an August 20 hearing at the Public Utility Commission of Texas as an important milestone. They didn't spell out what happens if the audit delays grid connection past their target dates: Sealy's first phase (209 MW) is targeted for mid-2027, Brazoria Phase I for year-end 2027.

The moratorium doesn't automatically void private contracts. But it does create timing risk on the Texas pipeline, which is where the valuation upside lives. If Texas grid approval slips by a year or more, the exclusive arrangements with the Sandersville tenant could lose their leverage. And a lower 14x multiple on delayed, unsigned deals becomes even less defensible.

The exception worth noting: data centers building on-site power generation that bypasses traditional grid connection are not covered by this freeze. Whether CleanSpark is pursuing that path for its Texas campuses is unclear.

The liquidity question

CleanSpark's balance sheet is the least dramatic part of the story. Cash sits at $202.6 million, with total liquidity of $917 million including $815 million in Bitcoin holdings (nearly 14,000 BTC). Long-term debt is $1.8 billion. The company needs this capital to fund the Sandersville build-out and hold through the mining downturn.

Management said on the call that they're "not ideological about the Bitcoin balance" and are willing to deploy holdings for accretive opportunities. That's a useful signal — it means the BTC stash isn't sacred, it's working capital. But it also means the Bitcoin position is a variable input, not a fixed asset. If crypto weakens further, the company may need to sell to fund construction, which would reduce the equity value B. Riley's model assumes they'll still hold.

The financial bridge

The Sandersville lease is the one signed piece that justifies looking at CleanSpark as a data-center infrastructure company. $330 million in average annual NOI, triple-net terms, first revenue in late 2027. If the market eventually prices that contracted cash flow at anywhere near the 20x multiple B. Riley applies, Sandersville alone supports a substantial portion of the $26 target.

But the stock won't get there on Georgia alone. Texas is the multiplier — 885 MW of potential capacity across Sealy and Brazoria that could add $10.9 billion in enterprise value. The catch is that Texas is regulatory risk, not contracted revenue. The exclusivity arrangement with the Sandersville tenant is a lead, not a commitment.

The tripwire is straightforward. If the Texas moratorium pushes grid approvals past mid-2027, the exclusive window with the tenant narrows. If CleanSpark can't convert the LOI into signed leases within a reasonable timeframe, the Texas valuation — and more than half of B. Riley's $26 target — evaporates.

The setup

The market is still pricing CleanSpark as a Bitcoin miner in distress. That story was accurate in Q3. But the Sandersville lease changes the trajectory: signed revenue, investment-grade tenant, triple-net structure, late-2027 start date. The numbers don't reflect that yet.

The question isn't whether Sandersville is real. It's whether Texas follows. If the regulatory freeze is short-lived and CleanSpark converts its exclusivity into signed leases, the $26 target has structural support. If Texas drags on, the stock stays anchored to mining economics and the SOTP model becomes an aspirational exercise.

When conviction is high and the financial bridge is explicit, you put a target and a tripwire on the thesis. The target is B. Riley's $26 — anchored in the Sandersville NOI multiple, conditional on Texas conversion within 12 months. The tripwire is a failed Texas regulatory outcome: if grid approvals are pushed past the mid-2027 Sealy delivery date without CleanSpark securing lease commitments, the Texas value proposition breaks.

Discipline over ego. The Sandersville deal is the proof point. Texas is the risk. The rest is noise.