CleanSpark Q2 Earnings 2026 Show $378M Loss as Bitcoin Fair-Value Swing Bites Miners

CleanSpark Inc. (NASDAQ: CLSK) reported a $378.3 million net loss for the quarter ended March 31, 2026, per the company’s official Q2 release. I’ll be blunt: that headline looks uglier than the operating story underneath it. This isn’t a broken-miner quarter. It is what happens when fair-value accounting drags the bitcoin treasury straight through the income statement. The $224.1 million non-cash hit on the BTC stack did most of the damage, while hashrate still climbed 18% year-over-year and megawatts under contract doubled. For CLSK holders and BTC traders, the read is pretty simple. The quarterly EPS line now follows the bitcoin chart almost line for line.

Revenue came in at $136.4 million, down $45.3 million (24.9%) from $181.7 million a year ago, according to the Q2 2026 financials. That decline tracks two concrete pressures: BTC price movement during the quarter and rising network difficulty across CleanSpark’s U.S. fleet. The reported loss of $1.52 per basic share compares with a $0.49 loss a year earlier. Cost of revenues was $81.7 million. Depreciation and amortization climbed to $115.9 million as the company kept buying rigs. Adjusted EBITDA, which strips out the bitcoin fair-value adjustment, printed at negative $241.2 million versus negative $57.8 million in the prior-year period. Not pretty. But not simple either.

Here’s the part I keep coming back to. CleanSpark held $260.3 million in cash and $925.2 million in bitcoin as of quarter-end, with the BTC stack up 14% year-over-year, per the balance sheet disclosures. Total assets stood at $2.9 billion against $1.79 billion in long-term debt and $986.2 million in stockholders’ equity. Working capital was $1 billion. Most earnings recaps will lead with the loss and move on. That’s only half right. The balance sheet is the real story, because CLSK now looks less like a plain bitcoin miner and more like a leveraged BTC vehicle with power assets bolted underneath it.

The adoption signal runs through the megawatts, not the EPS line. CleanSpark controls more than 1.8 gigawatts of power, land, and data center assets across the United States, with 585 MW of ERCOT-approved capacity in Texas, fresh ERCOT approval for 300 MW in Brazoria, and ongoing leasing in Georgia plus construction in Sandersville. Why does this matter? Because the same energy base can support bitcoin mining today and AI/HPC infrastructure later. That is the pitch. Counter to the usual “miners are just BTC beta” framing, CleanSpark is trying to make the power portfolio matter twice. Own the megawatts. Then find the highest-paying load.

CEO and Chairman Matt Schultz framed the quarter around four levers: land and power, leasing, financing, and construction. “This quarter, we accelerated our digital infrastructure evolution across four key areas: land and power development, with ERCOT approval of 300 MW in Brazoria; leasing, with further progress in Georgia and beyond; financing, as market conditions remain constructive; and construction, as we continue developing the new parcel in Sandersville,” he said in the company’s official statement.

“Our objectives are clear: commercialize our AI/HPC-applicable assets, grow the portfolio, and continue mining efficiently.”

President and CFO Gary Vecchiarelli called the balance sheet a competitive advantage heading into the next phase, saying CleanSpark ended the quarter with enough liquidity to support near-term execution while preserving optionality as AI and HPC infrastructure demand grows. The company also flagged uncertainty around potential tariff liability on miners purchased since 2024. Small line. Big risk. My take: that tariff note deserves more attention than it will get in most quick earnings summaries.

Worth pausing on. A $224.1 million non-cash mark-down on a $925.2 million BTC holding is roughly a 24% drawdown on the position over the quarter, almost exactly what spot BTC did in the period. That is not really a CleanSpark-specific failure. It is the mechanical effect public miners face when they hold treasury BTC under fair-value accounting, per FASB ASU 2023-08, which took effect for fiscal years beginning after December 15, 2024. Marathon, Riot, and Hut 8 will all be reading from the same chart when they report. Yes, this slightly contradicts the instinct to treat net income as the main scorecard. Bear with me. In this setup, CLSK shareholders should care less about the headline net-loss number and more about what moved underneath it: hashrate growth was up 18%, BTC stack growth was up 14%, and megawatts under contract doubled. Those went the right way.

What this means

The $378 million headline loss is mostly a non-cash accounting shock, not an operational collapse. Strip out the $224.1 million non-cash hit and the picture is a miner that grew hashrate, grew BTC reserves, and doubled contracted power into a falling tape. Is that enough to make the quarter good? No. But it does change what investors should actually judge. The signal for CLSK is that the equity will keep trading as a high-beta BTC proxy with an AI/HPC option attached. Until the company books its first material AI/HPC tenant revenue, the stock probably moves more with spot bitcoin than with any single fleet metric. For BTC itself, the read-through is cleaner: miners holding $925 million stacks are not selling here. CleanSpark’s BTC position grew 14% YoY through the drawdown.

Watch Q3 closely. First, the next ERCOT milestone on the Brazoria 300 MW matters because energization timing is the clearest catalyst for the AI/HPC narrative. Second, any update on the post-2024 miner tariff exposure CleanSpark flagged could move the next two quarters more than investors expect. Third, the next CME bitcoin futures basis and BTC spot level around the Sandersville construction updates will set the backdrop for the next print. We should stop pretending miner earnings now live in a separate lane from BTC marks. Every miner print from here will be benchmarked against the BTC mark on the reporting date, and CLSK reports next when the March-to-June quarter closes. The fair-value swing cuts both ways. If BTC marks higher into June 30, the same accounting rule that produced this $224 million loss will hand CleanSpark a paper profit just as large.