Cango Inc. (CANG) Q2 2026 Earnings Call Transcript

Cango reported Q2’26 total revenue of ~$50.8M, ~ $47.4M from Bitcoin mining, but posted a net loss of ~$81.6M driven largely by noncash impairment/disposal losses tied to restructuring of its mining asset base. The company said AI infrastructure progress and its first customer contract were signed after June 30, so they are not included in the quarter’s results. As of June 30, it held 1,056 Bitcoins and had ~$23M in cash/cash equivalents/cryptos versus ~$31.2M long-term debt.

Analysis

The market is likely to overreact to the AI contract narrative because it creates a new story before it creates new earnings. For now, this is still a levered BTC-linked equity with an unproven second engine, so any multiple expansion depends on recurring AI bookings and evidence that the business can be funded without dilution. Thin liquidity relative to debt means the pivot is only valuable if it quickly becomes self-financing.

The mining reset has two opposing effects. In the short run it should improve fleet quality and lower unit costs, but it also shrinks the revenue base and keeps reported results noisy with impairment and disposal losses. The second-order winners are stronger miners and infrastructure names that can absorb used ASICs or take share, while CANG’s hardware liquidation can temporarily pressure second-hand equipment pricing and benefit buyers with better balance sheets.

The consensus risk is that investors treat "AI" as a valuation shortcut instead of an execution problem. The next 1-3 months are about contract quality, gross margin, and whether the AI line item shows up as real revenue; over 6-18 months the key question is whether CANG becomes a durable infra operator or just a declining BTC proxy with optionality attached. Falsifiers are simple: no meaningful AI revenue by the next report, worsening financing needs, or a BTC drawdown that forces another restructuring.

Key Decisions for Investors

• Fade any sharp post-call rally in CANG with a small short or put spread; use it as an event-driven trade only if shares reprice on the AI story without disclosed revenue. Invalidate the short if the next update shows material AI backlog or first-quarter AI revenue at a scale that changes the mix.
• Preferred relative-value expression: long IREN or MARA versus short CANG to own the same crypto/AI theme through better-capitalized, more credible operators. This is a cleaner way to express skepticism on CANG’s pivot while keeping upside to BTC and AI infra spending.
• Watch for spillover into MARA, CLSK, and RIOT if CANG continues disposing of low-efficiency rigs; used-ASIC supply can temporarily soften hardware pricing and improve returns for miners buying cheap capacity. This is more of a 1-3 month tactical alert than a standalone trade.
• If you want BTC beta, use BTC or MSTR instead of CANG until AI revenue is visible and recurring. CANG’s equity is still a balance-sheet levered special situation, not yet a proven AI compounder.