From Automotive Finance to Bitcoin and AI Engines: An Analysis of Cango's 'What Not to Do' Strategy

Bitcoin mining company Cango rejects AI mega data centers, shifting towards distributed inference in the long-tail market.

Written by: Forbes

Compiled by: AididiaoJP, Foresight News

"From the outside, people must think this company is crazy," said Juliette. "Who are they? They make such bold moves while knowing nothing about this industry." She was referring to the day when a Chinese auto loan company invested hundreds of millions of dollars 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 rushing to lease power to large-scale cloud service providers building AI mega training clusters. However, 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 Chinese auto financing platform listed in the U.S. In November 2024, it agreed to acquire approximately 50 EH/s of mining machines from Bitmain, becoming a pure Bitcoin mining company. Subsequently, on April 13 of this year, it launched an AI inference subsidiary named EcoHash, equipped with its own software layer, EcoLink. There is no AI training and no new mega data centers being built. It is simply betting that the scattered small mining companies that large-scale cloud service providers cannot utilize are where a significant amount of AI computing power will land.

Note: 50 EH/s is a very large computing power scale. The total hash rate of the global Bitcoin network typically fluctuates around 600-800 EH/s, and 50 EH/s accounts for about 6-8% of the global total computing power, which is significant for a single acquisition by a large mining company, bringing substantial mining capacity.

"What not to do is as important as what to do," said Juliette, Cango's Senior Communications Director. She emphasized this phrase repeatedly. 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 wanted to own energy.

She is well aware of this history. She has worked at Cango for eight years, previously at The Wall Street Journal and consulting firm FTI. The story she tells begins with automobiles. Cango invested early in Chinese electric vehicle manufacturer Li Auto before its IPO. When Li Auto went public in 2020, Cango recorded a fair value gain of about 3.3 billion yuan (approximately $508 million) and became interested in the energy business behind the cars. By 2023, it began looking for energy projects in Australia and the Middle East.

"During a trip to the Middle East to look for solar projects, the management encountered Bitmain," Ms. Ye said. This is how the auto loan company met Bitcoin mining.

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

The entry cost is not cheap. Cango paid $256 million in cash in November 2024 to acquire 32 EH/s of mining machines from Bitmain, and then acquired another 18 EH/s through stock purchases from a company run by a former Bitmain CFO. To shed the 'Chinese concept stock' label, it sold its entire domestic auto business for about $352 million. It brought in a crypto-native leadership team, including a new CEO and a chairman who founded a financing company related to Bitmain, Antalpha. By mid-2025, the loan business will no longer exist. A mining company will take its place.

Why Everyone is Shifting

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

"The future of high-performance computing for AI may 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 hot commodity. "It's all an energy game," Wang said. "We believe that energy will become a more scarce asset for everyone in the future."

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

Timing and cycles align. "We have been very accurately following the four-year cycle," said crypto investor Michael Terpin on the On The Margin podcast. After each halving, mining profit margins tighten, 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 service provider CoreWeave, and other mining companies from IREN to the company formerly known as Bitfarms are also following suit. "Crypto mining warehouses are quietly shifting to AI inference, bringing in about four times the revenue," an analyst behind the @0xCristal account wrote on the X platform. "A GPU warehouse serving large language model inference earns more than mining blocks."

Betting Against Mega Sites

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

"We absolutely do not do AI training," Ms. Ye said. "That space is already crowded with large-scale 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 large-scale cloud service providers seeking 100 megawatt parks. But Ms. Ye believes this is perfect for the other half of AI. "For AI inference, you must deploy distributedly. You must be close to the customer to reduce latency," she said. "10 to 50 megawatts is too small for large-scale cloud service providers, but perfect for AI inference."

She then mentioned her favorite data. "Over 70% of the electricity in the mining industry is actually owned by individual players and small sites," Ms. Ye said. "Only 30% is controlled by those publicly listed mining companies." These small operators own land and electricity. They do not own AI technology, customers, or financing. Cango wants to bring all of this to them. "We provide them with a symbiotic relationship. We come to the site, bring AI, and they have land and electricity," 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 among these small sites."

EcoLink is the glue. A small site cannot compete with the always-on uptime of large-scale cloud service providers, so Cango spreads reliability. "If one side goes down, we can redirect workloads to another site in milliseconds," Ms. Ye said. The current buyers are what she calls long-tail customers. GPU leasing market platforms like Runpod and Vast.ai, distributed inference clouds like Zenlayer, and those AI startups too small to sign terms with large-scale cloud service providers. The price is the attraction: top providers may charge several dollars per GPU per hour, while the market rents the same chip for less than a dollar. Ms. Ye said 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 still operates about 31.7 EH/s, which brought in $98.4 million in mining revenue in the first quarter. This is the cash that keeps the company running while raising funds for AI. "Most mining companies have completely given up on Bitcoin mining," Ms. Ye said. "For us, it is more of a hybrid approach."

The cleanup has been brutal. "We are basically cleaning the deck," Ms. Ye said. "Investors may want to invest in our AI transformation, but they do not want their money used to pay off old debts." Therefore, Cango sold 6,451 Bitcoins for about $442 million and reduced its long-term debt from $557.6 million to $30.6 million in one quarter, a decrease of 94.5%. Its Bitcoin reserves have decreased to about 1,000 coins. Subsequently, it raised $75 million 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 last August. Ms. Ye calls it a "living showroom." Two to three more nodes will go online by the end of this year.

Skeptics

Not everyone is convinced. "People are a bit cautious about this," Wang said when discussing the AI frenzy, "because people are worried about a bubble." The story is ahead of revenue by years. Transforming a warehouse full of fans into a liquid-cooled AI data center is costly. Many mining companies saw their stock prices soar due to press releases, but gained nothing in reality. The company formerly known as Bitfarms saw its stock price rise hundreds of percent after its AI rebranding, but before earning a dollar in AI revenue. Analysts tracking these transformations continuously warn that the funds required to complete these transformations reach billions of dollars.

Bitcoin holders have different concerns. As miners shut down machines, the network's hash rate has declined, and some believe security costs have been overlooked. "Bitcoin miners are sacrificing the network for AI funding," warned a widely circulated post on the X platform. Cango's own buffer is thin. After the debt cleanup, it had only $7.2 million in cash at the end of the quarter, and at least one media outlet questioned its status on the NYSE. Even landmark deals are shaky: CoreWeave's $9 billion acquisition offer for Core Scientific fell through earlier this year.

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