Canaan Inc. CAN

Key data

The business

Canaan designs Bitcoin mining chips and builds them into machines sold under the Avalon brand. It shipped the world's first ASIC miners in 2013 and listed on Nasdaq in 2019. For a decade it made money one way: design a chip faster and more power-efficient than Bitmain's and MicroBT's, stamp it into a box, sell the box to mining farms. That business is brutally cyclical. Rig prices track the Bitcoin price with a lag, and when coins crash, farms stop buying and Canaan clears inventory below cost. Since 2023 it has bolted on a second business: running its own machines to mine Bitcoin and holding the coins. By mid-2026 it had 10.05 EH/s of installed computing power, mined 90 coins in June alone, and sat on a treasury of 1,867 Bitcoin and 3,952 Ether, powered at about $0.04 a kilowatt-hour across a West Texas footprint. The moat, such as it is, is chip design: efficiency per watt, with no software layer and no switching cost to lock a farm in. One overhang sits on top of it all. The stock has changed hands below $1 since January, and Nasdaq has put it on notice.

The numbers

The numbers tell two stories at once, a recovery in the yearly figures and a fresh air-pocket in the latest quarter.

Three of the last four quarters came in worse than analysts penciled in, the shortfall reaching six to seven cents a share in the two most recent. Q1 2026 revenue was $62.7 million, a fraction of the $529.7 million booked across all of 2025, the seasonal air-pocket that follows a strong Bitcoin-driven fourth quarter.

Step back and the arc is clearer. Revenue fell off a cliff from $772.8 million in 2021 to $211.5 million in 2023, then clawed back to $529.7 million. The line that matters most is gross profit: deeply negative in 2023 and 2024, when Canaan sold rigs for less than they cost to build, and positive again at $41.2 million in 2025. That is the recovery. The problem sits below it. Even with sales rebounding, 2025 still produced a $210 million net loss and $271 million of operating cash burn, the worst of the three-year stretch. Cumulative net losses since 2023 total $874 million, and cash has drained from $421 million at the end of 2021 to $43.5 million by March 2026. The gross-profit turn the March quarter hinged on has not printed again yet, because the June and August filings were monthly mining updates rather than a fresh income statement, so whether that positive number was the start of a trend or a single good quarter is still the open question. What the market looks to be underpricing is the repair in unit economics plus a marked treasury of nearly 1,900 coins, both ignored at two-thirds of book. The print that settles it is Q2 2026 gross profit and operating cash flow.

Management

Nangeng Zhang founded Canaan, runs it as chairman and chief executive, and in mid-June 2026 bought stock in the open market on three consecutive days. The dollar size was not disclosed in the filings, and no insider sold, which is the right direction even if a small signal. The board renewed a $30 million buyback in December 2025 and repurchased $5 million of stock during the year, then in August 2026 authorized management to fund further buybacks by selling down the Bitcoin treasury. That is a flexible idea that changes the optics without changing the arithmetic, since it swaps one asset for another below book value. Pay detail and the exact purchase dollars were not in the pulled data. The honest read: management is buying stock and shrinking the count at the same moment the company is burning cash, which only works if the cash outlasts the burn.

How it fails or surprises you

The cash runs out before the business turns. Canaan held $43.5 million of cash at March 2026 against $271 million of operating cash burn in 2025. Inventory, receivables and the coin treasury cushion that, but self-mining eats capital. If Q2 burn stays near the 2025 pace without a raise or a treasury sale, dilution or distress follows. The print: the Q2 cash balance and operating cash flow.

Revenue grows and cash still bleeds. In 2025 sales nearly doubled, yet the loss was $210 million and cash burn hit its worst level of the cycle. This is the fact my read explains least. If growth keeps failing to convert to cash, the gross-profit turn was noise. What would prove the recovery wrong: a Q2 gross margin back below zero, or operating cash burn no better than 2025's.

Bitcoin rips and the whole thing re-rates (right tail). At 0.33x sales the market pays nothing for a coin upcycle. If Bitcoin rallies, rig prices and order books jump, the self-mined coins and the 1,867-coin treasury mark up hard, and a sub-book miner can double or triple quickly. The market is not paying for this because it is pricing survival, not upside. The first tell: monthly Bitcoin production and treasury value climbing alongside a rebuilt Avalon order book.

Closing thoughts

This is an exposure, not a puzzle one print fully resolves. Canaan is a levered play on the Bitcoin price wearing a hardware company's clothes, and no earnings line changes that. The left tail is real and near: a company burning cash faster than it earns it, $43.5 million on hand, and a stock so low that Nasdaq is threatening to delist it, which likely forces a reverse split, optics rather than arithmetic. The right tail is equally real and fatter than the price implies, because at two-thirds of book and a third of sales you are handed the coin upcycle almost for free. What is at risk if the cash breaks is most of the equity. What the upside is worth if Bitcoin runs is a multiple of today's price. The fatter tail, in my judgment, is the right one, but only for capital that can survive the left.

The bet is still that Canaan keeps selling enough Bitcoin mining computers, and mining enough Bitcoin on its own, to stop burning through its cash before the cash runs out, and that Bitcoin stays expensive. What breaks it is simple to watch: the next cash balance against the next quarter's burn. If Q2 2026 shows gross profit holding positive and the operating cash drain narrowing, the recovery is real. If gross profit slips back under zero, the book is still melting, and no treasury maneuver fixes that.