Cango (CANG) Q2 2025 Earnings Call Transcript

Cango reported 50 exahash deployed (≈6% of global hashrate) and RMB 1.0 billion revenue in Q2 2025, with adjusted EBITDA of RMB 710.1 million but a net loss of RMB 2.1 billion driven by one-off discontinued operations and impairment. The company mined 1,404.4 BTC in the quarter (July 650.5 BTC), held RMB 843.8 million (~$118m) cash as of June 30, completed a $352m divestiture of China assets, and acquired a 50 MW Georgia site to lower power costs. Reported cash cost per mined BTC was ~$83,091 and all-in cost ~$98,636, while operational effective hashrate in August was ~44 EH (≈87% efficiency vs peers >90%), highlighting near-term cost and efficiency risks despite scale and an asset-light expansion strategy. Management will change reporting currency to USD starting Q3 2025 and continues to pursue M&A and energy + AI/HPC initiatives.

Analysis

Cango’s pivot creates a structural arbitrage between scale and capital intensity that the market is likely mispricing. An asset-light grower that outsources hosting and buys secondhand miners faces a different margin cadence than vertically integrated peers: lower depreciation but higher exposure to used-hardware pricing and hosting availability. That flips the usual valuation levers — short-term EBITDA sensitivity rises while durable free cash flow depends on sustained access to cheap, reliable power contracts and a stable secondary ASIC market.

Second-order supply effects matter. Continued appetite for turnkey racks will tighten the used-miner channel, pushing up replacement costs and shortening the economic life of low-efficiency fleets; vendors and resellers of used ASICs become de facto strategic counterparties. Grid constraints and seasonal curtailment in priority US regions introduce asymmetric downside in summer months, while permitting and conversion to HPC workloads create 6–24 month capex cliffs that can strand capacity or force equity raises if power or retrofit economics miss targets.

Key catalysts to watch are operational run-rate improvement over the next quarter, the integration timeline for newly acquired sites, and any discrete disclosures on long‑term power contracts or storage pilots. Positive readthroughs (faster-than-guided efficiency gains, below‑market power pricing) compress financing risk and should re-rate equity multiples quickly; conversely, upward pressure on network hash or spikes in used-rig prices would compress margins and reintroduce impairment risk within months.

From a competitive vantage, incumbents with deep balance sheets and vertical control will widen gaps on reliability and permitting for HPC conversion, but nimble buyers with capital can arbitrage the secondhand market in the short term. The stock is a binary operational story: execution against efficiency and power-cost targets drives outsized upside on relatively modest revenue moves, while any failure to close the operational gap or refinance at attractive rates creates rapid downside.

Key Decisions for Investors

• Long CANG equity sized to a tactical sleeve (5–7% portfolio crypto allocation) with a 3–9 month horizon — thesis: operational efficiency and low-cost site synergies re-rate EBITDA; set a stop-loss at 30% of position value and a target of +35–50% if quarterly operational metrics show sequential improvement.
• Relative-value pair: long CANG / short MARA (equal notional) for 3–6 months — trade the asset-light vs heavy-capex divergence. Reward: capture rerating if secondhand rig pricing tightens and Cango executes upgrades; risk: both fall if BTC or hash rate moves sharply against miners. Limit max drawdown to 25% per leg.
• Options collar to express asymmetric upside: buy CANG 9‑month call spread (buy nearer-term ATM, sell OTM to fund) sized to 2% notional and sell covered calls against the long equity sleeve if position outperforms — caps cost while leaving meaningful upside if operational catalysts hit in 2–4 quarters.
• Small, long-term satellite: buy NVDA (or NVDA LEAP calls) as a 12–24 month leash on the company’s stated AI/HPC pivot. Upside if conversion economics materialize; downside limited by allocating only 1–2% portfolio weight given execution risk and long lead times.