Cango Inc enters AI market; capital restructuring demonstrates determination

With Bitcoin mining rewards halving, energy price fluctuations, and increasing competition, the global mining industry is undergoing a restructuring. Diversifying business lines and establishing sustainable revenue models have become common challenges for leading companies in the sector. Among these, transforming existing computing infrastructure into artificial intelligence (AI) computing services has emerged as a notable direction of change. Recently, publicly traded mining company CANG Group (CANG.US) explicitly outlined its AI infrastructure development roadmap in its shareholder letter.

From Mining Network to AI Node Network

The reusability of computing resources is a core technological driver. Whether it’s the mining machines required for Bitcoin mining or GPU clusters used for AI training and inference, both are fundamentally large-scale parallel computing units. They share similarities in cabinet deployment, cooling management, and network maintenance. The operational experience accumulated by mining companies managing centralized computing centers provides a solid foundation for managing AI computing facilities.

Furthermore, the advantage of energy resources creates an insurmountable competitive barrier. AI computing is often called a “power-consuming beast,” with its development speed also limited by the stability and economic viability of energy supply. Large mining companies like CANG have deployed low-cost, diversified energy infrastructure worldwide, solving complex issues related to grid connection and load balancing. As mentioned in CANG’s shareholder letter, there is an energy gap in the AI era, and their global grid-connected infrastructure is key to seizing opportunities, enabling them to offer AI computing power at marginal energy costs lower than traditional data centers.

Empowering Long-Tail Miners to Co-Create AI Infrastructure

CANG’s competitive advantage does not solely come from its own energy or computing resources but from its ability to access a global network of small and medium-sized mining farms. These farms are located in regions with vastly different energy prices and supply-demand structures. Originally serving crypto mining, these farms are now being integrated into a distributed infrastructure system capable of supporting AI computation through platform deployment. According to public information, CANG has established a wholly owned subsidiary, EcoHash Technology, focused on AI computing, and appointed an AI Chief Technology Officer to lead a dedicated team for technical execution.

Traditional AI infrastructure is mostly concentrated in ultra-large cloud or data centers, with high entry barriers. CANG’s approach is different: by offering lightweight, modular GPU solutions, it enables small and medium miners to participate in the AI computing market at low cost. For these miners, underutilized scattered energy resources are transformed into stable productive capacity for AI computation through intelligent scheduling.

For the industry as a whole, this not only expands the geographic coverage of AI computing power but also creates a decentralized, more energy-efficient infrastructure layer. CANG’s role thus shifts from merely providing energy and computing power to becoming an AI engine for the global long-tail mining ecosystem.

Short-Term Monetization and Long-Term Vision

CANG’s AI transformation roadmap demonstrates a clear strategic progression. It is divided into three phases: in the near term, the group will deploy modular, containerized GPU computing nodes as a foothold in the market. This “plug-and-play” solution can be rapidly deployed on the group’s existing global infrastructure, aiming to meet the massive long-tail AI inference demands of small and medium-sized enterprises.

In the mid-term, CANG plans to develop proprietary software-defined orchestration platforms to integrate dispersed physical computing nodes worldwide into a unified, flexible, enterprise-grade computing network. This step is crucial for transitioning into a platform operator, reducing technical barriers for customers to utilize distributed computing resources.

Looking further ahead, the goal is to build a mature global AI infrastructure platform that not only mobilizes idle energy within its own mining ecosystem but also integrates broader underutilized power resources. Ultimately, revenue streams will be established through platform services and computing protocols that span market cycles.

Reallocating Capital to Empower the Computing Landscape

Financial maneuvers during the transformation, especially the adjustment of Bitcoin holdings, often attract market attention. These should be understood from a strategic perspective. CANG’s sale of some Bitcoin assets to strengthen its balance sheet and reduce leverage aims to raise funds for expanding AI computing infrastructure. This is a rational reallocation of financial resources—converting high-volatility assets into productive capital capable of generating future cash flows. Similar strategies are seen in companies like CleanSpark and Marathon, reflecting a balancing act between “capitalizing on crypto asset upside potential” and “investing in future certainty.”

More recently, CANG’s capital actions further demonstrate its commitment to transformation. EWCL completed a $10.5 million equity investment, and CANG’s chairman and wholly owned entity signed an agreement to invest up to $65 million in equity, with proceeds explicitly allocated to support AI and computing infrastructure expansion, further strengthening the company’s balance sheet. These capital reallocations are not only substantive endorsements of the company’s strategic direction but also reflect management’s long-term confidence in the AI infrastructure track.

Diversification of business lines also serves as a financial hedge against cyclical volatility. The high volatility of the crypto market causes mining revenues to fluctuate with coin prices, whereas AI computing services that generate recurring cash flow can smooth earnings and enhance the company’s valuation in capital markets, supporting more stable long-term growth.

CANG’s approach reveals another possibility for miners transitioning into AI: seeking efficiency advantages at the intersection of energy and computing power. Through platform-based integration and distributed deployment, CANG reconfigures power, land, and cooling assets across its global mining network, continuously contributing value in the AI domain. This model is not only cost-effective but also has evolutionary potential—it can dynamically switch between crypto computing, AI inference, and local computing tasks in response to market and technological changes, maximizing resource utilization.