Cango (CANG) Q2 2026 Earnings Call Transcript

Cango reported Q2 2026 revenue of $50.8 million, down approximately 50% sequentially, and an $81.6 million net loss, driven largely by $42.9 million of mining-machine impairment and $8.5 million of disposal losses. Bitcoin mining revenue was $47.4 million on production of 656 BTC, while cash mining cost fell 5% sequentially to $73,313 per BTC as the company reduced hashrate and shifted capacity toward leasing. Cango completed a 3MW Georgia AI infrastructure site in July and signed its first customer contract, with modest AI revenue expected to begin in Q3; however, possible summer power curtailments and continued mining-price exposure remain risks.

Analysis

CANG’s core issue is not the impairment charge but the residual earnings power after the fleet reset: cash mining cost remains close enough to Bitcoin’s marginal economics that difficulty increases or summer curtailments can quickly erase any apparent improvement. Leasing reduces operating-cost volatility but also converts upside exposure into lower-quality, less transparent revenue; the market should assign a lower multiple until management discloses lease economics, contracted duration, and cash conversion. The BTC-denominated borrowing hedge limits downside over roughly one-to-two months of production, but it also caps participation in a sharp BTC rally and introduces collateral/refinancing risk in a volatile asset.

The AI optionality is likely being overcapitalized by retail investors. A 3MW facility is commercially immaterial versus public GPU-hosting peers such as CoreWeave or larger power-backed infrastructure operators, and there is no disclosed customer term, utilization commitment, GPU ownership model, pricing, or incremental capex. The more important read-through is negative for conversion ambitions: most of CANG’s mining footprint is third-party hosted, limiting its ability to repurpose power contracts into durable AI capacity; a successful Georgia pilot does not establish scalable site control.

Near term, Q3 can look optically better as impairment rolls off and AI revenue begins, but curtailment plus still-negative adjusted EBITDA should keep liquidity central. Over 6-18 months, CANG requires either materially lower cash cost, sustained BTC appreciation without hedge drag, or externally financed AI expansion; each path risks dilution or increased related-party leverage. This is a balance-sheet and execution story, not a clean AI rerating.

Contrarian upside exists only if the first AI contract demonstrates high utilization and positive site-level contribution margins, allowing a credible valuation of the Georgia power asset above mining-equipment liquidation value. That thesis is falsified if Q3 cash burn persists, AI revenue remains de minimis, short-term BTC debt rises, or additional hardware impairments are required.

Key Decisions for Investors

• Avoid initiating CANG long exposure before Q3 disclosure. Require AI contract term, MW committed, utilization, customer concentration, incremental capex, and site-level EBITDA; absent these, treat AI revenue as an alert rather than a valuation catalyst.
• Tactical short CANG or long BTC / short CANG pair over the next 1-3 months if the stock rallies on AI headlines without disclosed economics. The pair isolates CANG’s fleet, liquidity, and execution risk from broad Bitcoin beta; cover on verified AI utilization or a material cash-cost reduction below roughly $60,000/BTC.
• For Bitcoin upside, prefer liquid BTC exposure or higher-quality miners rather than CANG: CANG’s production upside is partly hedged and its leased capacity dilutes direct BTC sensitivity. Reassess only after a quarter of positive adjusted EBITDA excluding fair-value marks.
• Monitor cash plus unrestricted BTC versus debt and quarterly operating cash burn. A deterioration in net liquidity, expansion of BTC-denominated borrowing, or new related-party financing would be a downside catalyst and warrants increasing any short exposure.