Cango Georgia bitcoin mining site fire injures two workers
An October fire at Cango's bitcoin mining facility in Soperton, Georgia leaves two workers injured and raises questions about potential disruption at the site. The incident occurs at a location that also supports the company's expansion beyond crypto mining into AI computing services.
Highlights
- Cango's bitcoin mining facility in Soperton, Georgia, suffers a fire on October 22 injuring two workers, with the cause under investigation.
- Cango acquires the 50-megawatt Soperton site for $19.5 million in August 2025, allocating 30 megawatts to mining and 20 megawatts to hosting.
- EcoHash, a Cango subsidiary, launches commercial GPU computing at the site with up to 3 megawatts of infrastructure as of September 3.
Fire response and investigation at Soperton facility
As reported by The Energy Mag, citing WSB-TV, two workers are injured in a fire at Cango's bitcoin mining facility in Soperton and are airlifted to a burn unit in Augusta.The fire breaks out around 10:45 a.m. on October 22 at 881 Commerce Drive, citing the Treutlen County Emergency Management Agency. The Soperton-Treutlen Fire Department quickly extinguishes the blaze. Authorities do not release the workers' identities or the extent of their injuries when WSB-TV publishes its report on October 23, and the cause remains under investigation.
The address matches Data Center Map's listing for Cango's Soperton facility. Former owner LN Compute also identifies Soperton as the site it sells to Cango in August 2025.
Site role in mining and AI expansion
Cango announces the acquisition of the 50-megawatt Georgia facility for $19.5 million in cash on August 11, 2025. The company says the site had previously hosted its mining machines under a third-party agreement, and it plans to allocate 30 megawatts to its own mining operations and 20 megawatts to hosting other customers.The facility also supports Cango's push into artificial intelligence computing. Its EcoHash subsidiary says on September 3 that it begins commercial GPU computing services at the Georgia site after completing a dedicated section with infrastructure supporting up to 3 megawatts. The reports reviewed do not quantify equipment damage, operational downtime, or any effect on the AI computing section.
Our earlier coverage of the U.S. AI data center buildout highlighted mounting pressures on new projects, including higher financing costs, construction delays, and growing limits on available power. We noted that only a portion of planned developments are viewed as realistic, as investor sentiment weakens and communities and regulators increasingly push back on large-scale expansions.