Bitcoin mining company Cango rejects AI mega data centers and shifts to distributed inference long tail market.

Written by: Forbes

Translated by: AididiaoJP, Foresight News

“From the outside, people must think this company is crazy,” said Juliette. “Who are they? They made such a bold move without knowing anything about this industry.” She was referring to the day when a Chinese car loan company spent hundreds of millions to become a Bitcoin miner.

That was about a year and a half ago. Now, it is doing the opposite. Almost every publicly listed Bitcoin mining company is scrambling to lease power to the hyperscale cloud service providers building AI mega training clusters. But Cango (NYSE: CANG) is going against the tide.

Cango is currently in the third phase of its transformation. It went public in New York in 2018 as the only car financing platform from China listed in the US. In November 2024, it agreed to acquire about 50 Ehash mining machines from Bitmain, becoming a pure Bitcoin mining company. Subsequently, on April 13 this year, it launched an AI inference subsidiary named EcoHash, equipped with its own software layer EcoLink. No AI training, no new mega data centers. It is simply betting that the small fragmented mining companies that hyperscale cloud service providers cannot utilize will be where a large amount of AI computing power lands.

Note: 50 Ehash is an extremely large scale of computing power. The current total hash rate of the global Bitcoin network typically fluctuates around 600–800 EH/s, with 50 EH/s accounting for about 6–8% of the global total computing power, which is at the level of a large mining company's single acquisition and can bring significant mining capacity.

“What not to do is as important as what to do,” said Juliette, Cango's Senior Communications Director. She repeated this phrase multiple times. These nine words are the core of the entire strategy.

Energy first, Bitcoin second

Ms. Ye stated that the company never intended to mine Bitcoin from the beginning; it aims to own energy.

She is very familiar with this history. She has worked at Cango for eight years, previously employed by The Wall Street Journal and consulting firm FTI. The story she tells begins with cars. Cango invested early in Chinese electric vehicle manufacturer Li Auto before its IPO. When Li Auto went public in 2020, Cango recorded approximately 3.3 billion yuan (about 508 million USD) in fair value gains and developed an interest in the power business behind the vehicles. By 2023, it began scouting energy projects in Australia and the Middle East.

“During a trip to find solar energy projects in the Middle East, the management team unexpectedly met Bitmain,” said Ms. Ye. This is the encounter of a car loan company with Bitcoin mining.

What truly impressed them was not the coin, but the infrastructure. “All these mining sites are basically energy infrastructure,” said Ms. Ye. “The only reason for mining farms to exist is that they consume energy and convert energy into coins. We can still convert energy into other things.” Mining is just an entry point. “We never thought about doing Bitcoin mining from day one. We were thinking about operating energy infrastructure from day one.”

The entry cost is not cheap. Cango paid 256 million USD in cash in November 2024 to acquire 32 Ehash mining machines from Bitmain and then acquired another 18 Ehash through stock, which were given to a company run by former Bitmain CFO. To shake off the “Chinese concept stock” label, it sold its entire domestic automotive business for about 352 million USD. It brought in crypto-native leadership, including a new CEO and a chairman who founded a financing company related to Bitmain called Antalpha. By mid-2025, the loan business no longer exists. A mining company takes its place.

Why everyone is turning

Cango is not the only mining company transitioning for AI. The mathematics of mining and the mathematics of AI have intersected, both competing for the same thing: electricity.

“The future of high-performance computing for AI might be the past of Bitcoin mining,” said Leo Wang, an executive at Canaan Creative, on the On The Margin podcast. In 2021, miners were the villains, accused of consuming electricity. Now, the same electricity has become a scarce commodity. “This is all an energy game,” Wang said. “We believe energy will become a more scarce asset for everyone in the future.”

What miners hold, which AI labs crave, is not chips, but a plug. Building new substations and signing long-term grid contracts can take years. “When hyperscale cloud service providers look for suppliers that can provide short-term guaranteed power, they turn to Bitcoin miners because Bitcoin miners have invested capital and secured power,” Wang said. He added that miners are “lucky” that AI emerged just as block rewards are decreasing.

The timing and cycles are synced. “We have been accurately following the four-year cycle,” said crypto investor Michael Terpin on the On The Margin podcast. After each halving, mining profitability tightens, and operators look for a second way to make money.

The market has followed suit. Core Scientific was an early mover, leasing capacity to AI cloud provider CoreWeave, and other miners from IREN to the company once known as Bitfarms are also following suit. “Crypto mining warehouses are quietly shifting to AI inference, bringing in about four times the revenue,” wrote an analyst behind the @0xCristal account on the X platform. “A GPU warehouse serving large language model inference earns more than mining blocks.”

Betting against giant sites

This is where Cango stands out. The popular practice is to transform a few large sites into AI training parks and sign long-term lease agreements with a hyperscale cloud service provider. Cango rejected this approach.

“We absolutely do not do AI training,” Ms. Ye said. “That space is saturated with hyperscale cloud service providers. Competing with them is unrealistic.” This decision stems from the company’s own scale. Cango has over 30 sites globally, most ranging from 10 to 50 megawatts. Too small to meet the demands of hyperscale cloud providers that seek 100 megawatt sites. But Ms. Ye believes this is perfect for another aspect of AI. “For AI inference, you must deploy distributed systems. You must be close to the customer to reduce latency,” she said. “10 to 50 megawatts is too small for hyperscale cloud providers, but perfect for AI inference.”

She then mentioned her favorite data point. “Over 70% of the power in the mining industry is actually owned by individual players and small sites,” said Ms. Ye. “Only 30% is controlled by those listed mining companies.” These small operators own land and power. They do not possess AI technology, customers, or financing. Cango wants to bring all this to them. “We provide them with a symbiotic relationship. We come to the sites, bring AI, they have land and power,” she said. “If there is anything that can help Cango establish a foothold in AI in the next three to five years, it is this symbiotic relationship between small sites.”

EcoLink is the adhesive. A small site cannot compete with the always-online uptime of hyperscale cloud providers, so Cango diversifies reliability. “If one side goes down, we can redirect the workload to another site in milliseconds,” said Ms. Ye. The buyers so far are what she calls long-tail customers. GPU rental market platforms like Runpod and Vast.ai, distributed inference clouds like Zenlayer, and those AI startups that are too small to sign terms with hyperscale cloud providers. The price is where the attraction lies: top providers may charge several dollars per GPU per hour, while the market rents the same chip for less than one dollar. Ms. Ye mentioned that there are no exclusive agreements with early test customers, and most have renewed. “Customer demand is absolutely real.”

Cash engine and costs

Cango has not given up on Bitcoin. It is still operating about 31.7 Ehash, bringing in 98.4 million USD in mining revenue in the first quarter. This is the cash that the company maintains operations with while raising funds for AI. “Most mining companies have just totally abandoned Bitcoin mining,” Ms. Ye said. “For us, it’s more of a hybrid approach.”

The cleanup has been harsh. “We are basically clearing the deck,” Ms. Ye said. “Investors may want to invest in our AI transformation, but they do not want their money to be used to pay old debts.” Therefore, Cango sold 6,451 Bitcoins, amounting to about 442 million USD, and reduced long-term debt from 557.6 million USD to 30.6 million USD in one quarter, a decrease of 94.5%. Its Bitcoin reserves decreased to about 1,000 coins. Subsequently, it raised 75 million USD for the launch of EcoHash. The first AI node will be deployed at a 50-megawatt site in Georgia that Cango acquired for 19.5 million USD last August. Ms. Ye calls it a “living showroom.” Two to three more nodes will go live by the end of this year.

Doubters

Not everyone is convinced. “People are a bit cautious about this,” Wang said when discussing the AI craze, “because people are worried about a bubble.” The story leads revenue by years. Converting a warehouse filled with fans into a liquid-cooled AI data center is expensive. Many mining companies saw their stock prices soar due to press releases but made no real gains. The company once known as Bitfarms saw its stock price increase by hundreds of percent after its AI renaming but before earning a dollar in AI revenue. Analysts tracking these transitions constantly warn that the funds required to complete these transformations amount to billions.

Bitcoin holders have different concerns. As miners shut down machines, the network's hash rate has declined, and some believe the cost of security has been overlooked. “Bitcoin miners are sacrificing the network for AI funding,” warned a widely circulated post on the X platform. Cango itself has thin buffers. After the debt cleanup, it had only 7.2 million USD in cash at the end of the quarter, and at least one media questioned its status on the NYSE. Even landmark deals are swinging: CoreWeave's 9 billion USD acquisition bid for Core Scientific was already canceled earlier this year.

Ms. Ye's response is discipline, which runs through everything she says. Mega sites and landmark training leases will be owned by the giants. Cango is betting on the rest: the thousands of megawatts of electricity spread among small independent miners, as well as the electricity that giants find difficult to easily access. She believes a significant amount of AI inference will quietly operate there.