NextFin News - Canaan Inc. reported second-quarter revenue of US$31.9 million, down 68.2% from a year earlier, and a net loss of US$97.6 million, as a weaker bitcoin price crushed both the demand for its mining rigs and the value of the bitcoin it mines itself. The Singapore-based company, listed on Nasdaq as CAN, said the three months ended June 30 combined "renewed bitcoin price pressure, weaker mining economics, and seasonal power constraints," and investors reacted immediately: shares fell as much as 18.7% intraday before settling roughly 3% lower near $0.35, on trading volume about 40 times the daily average.
The loss more than doubled from the US$88.7 million reported in the first quarter, while revenue nearly halved from US$62.7 million. Just twelve months earlier, in the second quarter of 2025, Canaan posted revenue of US$100.2 million and a net loss of only US$11.1 million. The contrast captures what has changed in the bitcoin-mining complex: the same company that called the second quarter of 2025 its strongest performance in ten quarters is now leaning on asset sales and share buybacks to preserve per-share value.
Revenue Collapsed on Two Fronts
Canaan has two engines, and both stalled at once. Product revenue - the Avalon mining machines it sells to other miners - fell to US$13.6 million from US$42.9 million in the first quarter and US$71.9 million a year earlier. Mining revenue - the bitcoin it produces with its own fleet - slipped to US$17.7 million from US$19.1 million sequentially and US$28.1 million in the year-ago quarter. Cost of revenues ran to US$61.2 million, leaving a gross loss of US$29.3 million, wider than the US$22.9 million gross loss in the first quarter and a sharp reversal from the roughly break-even gross result in the second quarter of 2025.
Q2 2026 presented a difficult period for bitcoin mining, as renewed bitcoin price pressure, weaker mining economics, and seasonal power constraints weighed on equipment demand and profitability.
Nangeng Zhang, Canaan's chairman and chief executive officer, said in the company's earnings release. Chief financial officer Jin "James" Cheng added that product revenue reflected "softer demand for mining rigs and a lower average selling price for computing power," while mining revenue suffered "amidst bitcoin price pressure and seasonal curtailments."
The revenue miss was not marginal. Consensus among analysts tracking the print stood at roughly US$38.6 million; Canaan delivered US$31.9 million, a shortfall of about US$6.8 million. On a per-share basis, the GAAP loss came to US$0.90 per ADS, compared with an expected loss of about US$0.06 - a miss of US$0.84. When a company misses on the top line by nearly a fifth and on earnings by more than tenfold, the market reprices the story, and Canaan's did: after an 18.7% trough, the stock settled early trade down about 3% at $0.35, with a market capitalization near US$260 million.
The Loss Was Mostly Accounting - But That Is the Point
The headline loss of US$97.6 million was dominated by non-cash charges that flowed directly from the decline in cryptocurrency prices during the quarter: a US$25.3 million write-down of inventory and prepayments plus reserves for inventory purchase commitments, a US$9.2 million impairment of property and equipment, and a US$18.2 million loss from fair-value changes in cryptocurrency holdings.
These accounting charges weighed on our bottom line but did not alter our focus on cost control, liquidity, and operational efficiency.
Cheng said in the release. That framing is technically correct and economically incomplete. A US$18.2 million fair-value loss on bitcoin holdings is not a cash outflow in the quarter, but it is a mark of an asset whose market price fell. A US$25.3 million inventory write-down is also non-cash, but it is a confession that the company paid more to build machines than the market is now willing to pay for them. The write-down is the accounting translation of a real economic fact: the cost embedded in the product exceeded its realizable value. Non-GAAP adjusted EBITDA, which strips out these items, still came in at a loss of US$74.9 million, compared with a positive US$25.3 million in the second quarter of 2025.
In other words, even after removing every non-cash charge, the business lost money - because revenue collapsed faster than management could cut costs. The cash picture shows the strain: cash and equivalents fell to US$66.0 million at June 30 from US$80.8 million at the end of 2025. In late August, after the quarter closed, Canaan monetized part of its digital-asset treasury, selling 3,952 ETH and 54 BTC for approximately US$13.9 million in cash. As of September 8, it had repurchased about 16.4 million ADSs for US$7.4 million under its US$30 million buyback authorization.
The Mechanism: Why a Bitcoin Price Slide Hits Miners Twice
The transmission channel from a weaker bitcoin price to Canaan's income statement runs through two distinct doors, and understanding both explains why the damage compounds rather than merely adds up.
The first door is the mining operation itself. Canaan produced 243 bitcoins in the quarter from an installed non-joint-venture fleet of 10.05 EH/s, up 23.3% year over year. More hash rate should mean more bitcoin and more revenue. It did not, because the dollar value of each coin fell. Bitcoin tested the US$60,000 support zone repeatedly in June, with intraday lows briefly approaching US$58,000, and closed the quarter down roughly 12% from the first quarter. American Bitcoin, a peer, reported revenue per bitcoin mined of about US$71,900 in the second quarter, down roughly 5% sequentially, and described the period as a roughly 12% quarter-over-quarter bitcoin price decline. Canaan's mining revenue of US$17.7 million, down from US$19.1 million despite a larger fleet, is the same story in miniature: production held up, pricing did not.
The second door is the equipment business, and it is where the leverage bites hardest. Canaan's customers are other miners. When bitcoin falls, their expected revenue per coin falls, their payback periods stretch, and their incentive to buy new machines evaporates. That is why product revenue fell 68% year over year, from US$71.9 million to US$13.6 million. The equipment maker is a leveraged bet on the mining cycle: a 12% decline in the bitcoin price can translate into a 60% or larger decline in machine orders, because capital expenditure is the first line item miners cut.
The third effect is the balance-sheet mark. Canaan held a record 1,915.5 BTC and 3,951.7 ETH at quarter-end. A larger treasury is a strategic asset when bitcoin rises; when it falls, it becomes a source of quarterly fair-value losses and, eventually, a source of liquidity. The company's stated plan - to "monetize a portion of our digital asset treasury to fund stock repurchases" - is a direct acknowledgment of that duality.
We believe the Company's current equity valuation understates the combination of our digital asset treasury, cash liquidity, and our mining operations, which continue to generate bitcoin rewards.
Zhang said, laying out the bull case in a single sentence.
This Is Cyclical - But the Cycle Has Sharper Teeth
The central question for investors is whether this is a cyclical downturn that will revert or a structural break in the business model. The evidence points to cyclical, with a structural complication attached.
The cyclical case is strong. Bitcoin mining is a commodity business with a long history of mean reversion. When prices fall, marginal hash rate shuts down, mining difficulty adjusts lower, and the surviving miners capture a larger share of block rewards at lower cost. That adjustment was already visible in 2026: bitcoin's mining difficulty dropped 10.09% in a single adjustment - from 138.96 trillion to 124.93 trillion - the second-largest negative adjustment of the year, after an 11.16% drop in February. The average network hash rate fell to between 740 and 886 EH/s, down 12% to 23% from October peaks. Canaan's own all-in power cost of roughly US$0.043 per kilowatt-hour is among the lowest in the industry, which is exactly the profile of a miner built to survive a downturn rather than be wiped out by one.
History supports the reversion view. Canaan's second-quarter 2025 revenue of US$100.2 million came just four quarters after the depths of the prior downturn; the swing from loss to profit and back again is the normal rhythm of a bitcoin-exposed hardware and mining business. If bitcoin stabilizes or recovers, equipment demand returns, inventory write-downs reverse, and the treasury becomes an asset again instead of a drag.
The structural complication is that the cycle's amplitude is increasing. The post-halving economics of 2024 cut block rewards to 3.125 BTC, and network difficulty reached record highs in late 2025 before rolling over. That combination means the margin for error is thinner than in previous cycles: miners need cheaper power, more efficient machines, and deeper balance sheets to survive the trough. Canaan's US$25.3 million inventory write-down is a signal that even a well-positioned miner can be caught holding product whose economics changed faster than it could sell. The cycle still reverts - but each trough claims more casualties, and the survivors need more capital to reach the next peak.
The Counter-Thesis: What If the Treasury Is the Trap, Not the Floor?
The bullish reading of Canaan rests on sum-of-the-parts arithmetic: the digital-asset treasury plus cash plus the mining operation should be worth more than a US$260 million market capitalization. The strongest argument against that view is that the treasury is not a passive floor - it is a funding source that management is already drawing down, and each draw reduces the very asset that supposedly backs the valuation.
This is not a fringe concern. In the second quarter, the broader mining sector showed the same pressure in sharper relief. Riot Platforms mined 1,587 bitcoins, an 11% increase year over year, yet bitcoin mining revenue fell 19% to US$113.7 million because the average price per coin fell. Riot's all-in cost to mine one bitcoin reached US$90,631, or 126.5% of the US$71,667 average production value of each coin - meaning the company lost cash on every bitcoin it mined before any corporate overhead. Riot posted a net loss of US$237.2 million, compared with net income of US$219.5 million a year earlier. If a larger, more diversified operator is mining at a cash cost above the market price, Canaan's smaller scale offers no insulation.
The counter-thesis, then, is that Canaan's treasury-driven buyback is not value creation but a slower liquidation: selling ETH and BTC to repurchase ADSs supports the share price today while shrinking the asset base that underwrites it tomorrow. The falsifying signal for the cyclical-reversion view is specific and observable: if bitcoin trades below US$60,000 for a sustained period - say, through the end of 2026 - and Canaan's mining cash contribution before depreciation turns negative for two consecutive quarters, then the downturn has moved from cyclical to structural for this balance sheet, and the treasury becomes a dwindling runway rather than a strategic reserve.
What to Watch: Three Horizons
Short term (sentiment and liquidity): The next catalyst is the earnings call and any update on the buyback pace. With US$7.4 million deployed of the US$30 million authorization and roughly US$13.9 million raised from the late-August treasury sale, management has both the mandate and the cash to continue supporting the shares. The watch item is whether repurchases accelerate or stall as the stock moves.
Medium term (fundamentals): The third quarter will show whether the difficulty reset that began in 2026 translates into improved mining economics. Key metrics: bitcoin production versus the 243 BTC of the second quarter, the average realized price per bitcoin, and whether product revenue stabilizes above US$20 million as customers restock. A recovery in computing-power sales - the company sold 2.5 EH/s in the second quarter, compared with 6.4 million TH/s in the strong second quarter of 2025 - would be the first sign that equipment demand is thawing.
Long term (structure): Canaan is attempting to move beyond pure mining hardware into compute-plus-energy infrastructure, including its 49% stake in Project ABC, which reached 4.85 EH/s of installed capacity by the end of July, and new Avalon Home products for household heating. The long-term question is whether these initiatives diversify revenue enough to reduce dependence on the bitcoin cycle, or whether they remain small appendages to a business whose fate is still set by the price of one asset.
Base case: bitcoin stabilizes in the US$60,000-US$70,000 range, difficulty continues to adjust lower, and Canaan's revenue recovers modestly in the second half of 2026 as equipment demand returns and the treasury provides a liquidity cushion. Upside case: a bitcoin rally above US$80,000 reignites miner capital expenditure, inventory write-downs reverse, and the sum-of-the-parts valuation re-rates higher. Downside case: bitcoin holds below US$60,000, mining cash contribution turns negative, and the company is forced to monetize more of the treasury - or issue new shares - to fund operations.
The takeaway is not that Canaan is broken; it is that the bitcoin cycle is doing what it always does, only with less room for error. Canaan's low power cost and record treasury give it a credible path through the trough - but the same assets that look like a floor at US$70,000 bitcoin can look like a shrinking runway at US$55,000, and the market is currently pricing the latter.
Explore more exclusive insights at nextfin.ai.
