Canaan (CAN) Q2 2026 Earnings Call Transcript

Canaan reported Q2 2026 revenue of $31.9 million, down from $100.2 million a year earlier and below its $35 million-$45 million guidance, while net loss widened to $97.6 million amid a $25.3 million inventory write-down, $9.2 million equipment impairment and $9.3 million digital-asset fair-value loss. Weak bitcoin prices, cautious miner procurement and industry inventory drove product revenue to $13.6 million and pressured ASP to $5.50/TH; management guided Q3 revenue down to just $11 million-$15 million. The company improved cash to $66.0 million from $43.0 million, holds digital assets with a $112 million carrying value, and repurchased 16.4 million ADSs for $7.4 million, but remains subject to Nasdaq minimum-bid compliance requirements through January 11, 2027.

Analysis

CAN is shifting from a hardware vendor into a leveraged mix of BTC treasury, self-mining and speculative power-development optionality, but the near-term setup is worse than headline impairments imply: a sharply lower next-quarter revenue run-rate leaves fixed R&D and corporate costs underabsorbed. Mining’s positive contribution margin does not establish economic profitability after depreciation, fleet-refresh capex and central costs; absent a material hash-price recovery, cash can erode despite reported liquidity.

The buyback is not a clean valuation signal. Selling volatile treasury assets to repurchase a sub-$1 ADS increases equity beta to both BTC and listing compliance, while reducing the asset cushion supporting operations. A reverse split remains a realistic compliance path; it is mechanically neutral but often pressures microcap liquidity and can reopen ATM-financing risk. The key 1-3 month catalyst is whether equipment pricing stabilizes before inventory is marked down again; management’s power-pipeline claims have no disclosed sites, counterparties, contracted load, customer commitments, or required capex, so the AI/HPC narrative should receive little valuation credit until independently documented.

CIFR is the cleaner relative beneficiary of investor demand for power-constrained compute infrastructure: it has a more credible route to monetize data-center capacity, whereas CAN would need external capital and partners before converting power discussions into HPC revenue. Contrarian upside in CAN requires BTC/hash price to rise enough to clear industry inventories, enabling operating leverage in both miner sales and treasury value; however, that same macro improvement is likely better expressed through BTC miners with funded infrastructure rather than CAN’s execution-heavy equity.

Key Decisions for Investors

• Maintain/establish short CAN versus long CIFR over the next 1-3 months; use a 20-25% stop on the CAN leg. Thesis: CAN’s hardware-price and funding risk should persist while CIFR better captures validated power/HPC optionality.
• Do not buy CAN solely on the repurchase program. Reassess only after disclosure of signed power capacity, counterparty, committed capex/funding source, and a credible path to listing compliance; absent these, treat the power pipeline as an alert rather than an investable catalyst.
• For crypto exposure, prefer liquid BTC or a basket of scaled miners over CAN through year-end. A sustained hash-price recovery and sequential product-revenue stabilization would falsify the bearish CAN operating thesis.
• Monitor CAN’s cash balance net of restricted cash, additional digital-asset sales, inventory turns, and any ATM use at the next filing. Material cash burn, a reverse-split announcement, or renewed equity issuance strengthens the short thesis; no new write-downs plus improving ASPs would require covering.