From Auto Finance to Bitcoin to AI — Kongge’s “Things You Shouldn’t Do”

Author: Forbes; Compiled by AididiaoJP, Foresight News

“From the outside, people are bound to think this company is crazy,” Juliet said. “Who are they? They’ve made such bold moves, yet they know nothing about this industry.” She was referring to the day a Chinese auto-loan company spent several hundred million dollars to become a Bitcoin miner.

That was about a year and a half ago. Now it is doing the opposite. Nearly every listed Bitcoin mining company is scrambling to lease power to the mega-scale cloud service providers that are building massive AI training clusters. Yet Cango (NYSE ticker code: CANG) is going the other way.

Cango is currently in the third stage of its transformation. It went public in New York in 2018, at a time when it was the only Chinese auto financing platform listed in the United States. In November 2024, it agreed to acquire Bitmain mining rigs worth about 50 exahashes, becoming a pure-play Bitcoin mining company. Then, this April 13, it launched an AI inference subsidiary called EcoHash, equipped with its own software layer, EcoLink. No AI training, and no newly built massive data centers. It is simply betting that those fragmented small mining firms—ones the mega-scale cloud service providers cannot use—are exactly where a large share of AI compute will land.

Note: 50 exahashes is an extremely large scale of computing. The global Bitcoin network’s total hashrate typically fluctuates around 600–800 EH/s, and 50 EH/s accounts for about 6–8% of the global total hashrate. That is on the scale of a one-off acquisition by a major mining firm, and it can deliver significant mining capacity.

“What shouldn’t be done is as important as what should be done,” said Juliet, Cango’s senior director of communications. She repeatedly emphasized this line. Those nine words are at the heart of the entire strategy.

Energy first, Bitcoin second

Ms. Ye said the company never planned to mine Bitcoin from the start—it wanted to own the energy.

She knows this history inside out. She has worked at Cango for eight years, and previously worked at The Wall Street Journal and the consulting firm FTI. Her story starts with cars. Cango took an early stake in China’s electric vehicle maker Li Auto, before Li Auto’s listing. When Li Auto went public in 2020, Cango recorded a fair value gain of about RMB 3.3 billion (about $508 million), and it became interested in the power business behind the cars. By 2023, it began scouting energy projects in Australia and the Middle East.

“On a trip to look for solar projects in the Middle East, management happened to meet Bitmain,” Ms. Ye said. That is how the auto-lending company crossed paths with Bitcoin mining.

What truly moved them wasn’t the coins, but the lines. “All of 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 was just an entry point. “From day one, we weren’t thinking about Bitcoin mining. What we wanted from day one was to operate energy infrastructure.”

The entry cost is steep. In November 2024, Cango paid $256 million in cash to acquire Bitmain’s 32 exahash mining rigs, and then—through a stock deal—acquired another 18 exahashes, with those shares going to companies run by former Bitmain financial executives. To shed the “China concept stock” label, it sold its entire domestic auto business for about $352 million. It brought in crypto-native leadership, including the newly appointed CEO and a chairman who co-founded Antalpha, a financing company tied to Bitmain. By mid-2025, the lending business no longer existed. A mining firm took its place.

Why everyone is turning

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

“The future of high-performance computing for AI may be the past of Bitcoin mining,” Leo Wang, an executive at Jianke Yunzhi, said on the On The Margin podcast. In 2021, miners were the villains, accused of wasting electricity. Now the same electricity has become highly sought after. “It’s all an energy game,” Wang said. “We believe that in the future, energy will become a scarcer asset for everyone.”

What miners have is not what AI labs are really craving. They do not want chips; they want a plug. Building new substations and signing long-term grid contracts can take years. “When mega-scale cloud service providers look for suppliers that can guarantee power in the short term, they turn to Bitcoin mining companies, because Bitcoin mining firms have already invested and secured the power,” Wang said. He added that miners were “lucky” that AI arrived as block rewards were declining.

Timing and cycles match up. “We have been following the four-year cycle very accurately,” crypto investor Michael Terpin said on the On The Margin podcast. After each halving, mining profit margins tighten, and operators look for another way to make money.

The market has followed. Core Scientific was an early mover, leasing capacity to the AI cloud service provider CoreWeave; other mining companies—from IREN to companies previously known as Bitfarms—have also jumped on. “Crypto mining warehouses are quietly shifting to AI inference and bringing in about four times the revenue,” an analyst behind the @0xCristal account wrote on X. “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 approach is to turn a handful of large sites into AI training parks and sign long-term leases with mega-scale cloud service providers. Cango refused to do that.

“We absolutely do not do AI training,” Ms. Ye said. “That field is already crowded with mega-scale cloud service providers. It isn’t realistic for us to compete with them.” The decision comes from the company’s own scale. Cango has more than 30 sites globally, most in the 10 to 50 megawatt range. They are too small to satisfy mega-scale cloud service providers chasing 100 megawatt parks. But Ms. Ye believes that is precisely what the other half of AI needs. “For AI inference, you have to deploy in a distributed way. You have to be close to the customers to reduce latency,” she said. “10 to 50 megawatts is too small for mega-scale cloud service providers, but perfect for AI inference.”

Then she pointed to her favorite kind of data. “More than 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 listed mining companies.” These small operators own land and power. They do not own AI technology, customers, or financing. Cango wants to bring all of that to them. “We provide them with a symbiotic relationship. We come to the sites, bring AI, and they have the land and the power,” she said. “If there’s anything that will allow Cango to gain a foothold in AI over the next three to five years, it’s the symbiosis among these smaller sites.”

EcoLink is the glue. A small site cannot match the always-on uptime of mega-scale cloud service providers, so Cango distributes reliability. “If one side goes down, we can route workloads to another site within milliseconds,” Ms. Ye said. So far, the buyers are what she calls long-tail customers: GPU rental marketplaces such as Runpod and Vast.ai, distributed inference clouds such as Zenlayer, and AI startups whose scale is too small to sign with mega-scale cloud service providers on their terms. Price is the attraction: top-tier providers might charge a few dollars per GPU per hour, while the marketplace rents out the same chips for less than $1. Ms. Ye said there are no exclusive deals 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 running roughly 31.7 exahashes, which brought in $98.4 million in mining revenue in the first quarter. This is the cash the company uses to keep operating while raising funds for AI. “Most miners simply abandon Bitcoin mining entirely,” Ms. Ye said. “For us, it’s more of a hybrid approach.”

The cleanup is brutal. “We’re basically clearing the deck,” Ms. Ye said. “Investors might want to invest in our AI transformation, but they don’t want their money used to pay off old debts.” So Cango sold 6,451 Bitcoins—about $442 million—and reduced long-term debt from $557.6 million to $30.6 million within a quarter, a drop of 94.5%. Its Bitcoin reserves fell to about 1,000 Bitcoins. Then it raised $75 million to fund the launch of EcoHash. The first AI node will be deployed at the 50 megawatt Georgia site Cango acquired last August for $19.5 million. Ms. Ye calls it a “live showroom.” By the end of this year, two to three more nodes are expected to come online.

Skeptics

Not everyone buys into it. “People are a bit cautious about this,” Wang said about the AI boom, “because they worry about a bubble.” The story is running ahead of revenue by years. Turning a warehouse full of fans into a liquid-cooled AI data center is expensive. Many mining companies’ share prices have spiked on press releases but produced nothing in reality. The company once known as Bitfarms saw its stock rise by hundreds of percentage points after rebranding itself as AI—before it earned even $1 of AI revenue. Analysts tracking these transformations keep warning that the capital required to carry them out amounts to billions of dollars.

Bitcoin holders have different concerns. As miners shut down rigs, the network’s hashrate has fallen, and some people believe the cost of security is being overlooked. A widely circulated X post warned: “Bitcoin miners are giving up the network for AI funding.” Cango’s buffer is thin. After debt cleanup, cash at quarter-end was only $7.2 million, and at least one media outlet has questioned its standing on the NYSE. Even marquee deals are shaky: CoreWeave’s $9 billion acquisition offer for Core Scientific earlier this year has already fallen through.

Ms. Ye’s answer is the discipline permeating everything she said. Giant sites and marquee training leases will belong to the giants. Cango is betting on the rest: thousands of megawatts of power distributed among small independent miners, and the electricity that the giants can’t easily reach. She believes that a large share of AI inference will quietly run there.