Key Highlights
- Hyperscale Data (GPUS) has secured a major Master Services Agreement with a neocloud company based in California for AI-focused colocation at its Michigan data center
- The agreement spans 10 years with optional extensions, potentially delivering more than $1.2 billion in total revenue across a 20-year maximum period
- Contract includes expansion rights to 52 megawatts that could elevate overall deal value beyond $3.0 billion
- First phase covering 20 megawatts targets Q4 2026 for going live, with possible revenue generation beginning late September
- Michigan campus operations will transition from cryptocurrency mining toward AI computational infrastructure
On Tuesday, Hyperscale Data (GPUS) revealed it has entered into a comprehensive Master Services Agreement with a California neocloud enterprise to provide AI colocation and data center capabilities at its Michigan location.
Despite the substantial announcement, GPUS stock declined 10.59% during trading, closing at $0.2557.
The contract encompasses an initial rollout of 20 megawatts dedicated to AI computing infrastructure, scheduled to become operational during the fourth quarter of 2026. The MSA establishes a 10-year base term with provisions for two additional five-year periods.
Should the client utilize the complete 20-year duration, projections indicate the arrangement will produce revenue exceeding $1.2 billion.
Additionally, the agreement provides the client with expansion rights up to 52 megawatts in total capacity. A separate provision allows for an extra 32 megawatts, and if activated within the initial two-year window and maintained throughout both extension periods, total contract revenue would surpass $3.0 billion.
Alliance Cloud Services, functioning as an indirect wholly owned subsidiary of Hyperscale Data, is undertaking retrofitting work on approximately 60,000 square feet at the Michigan campus to accommodate the client’s requirements. Capital expenditure estimates for the first 20-megawatt phase range from $100 million to $120 million.
CEO William Horne characterized the Michigan campus as being “positioned to offer a top-tier AI compute environment,” noting that service delivery could commence with revenue generation “as soon as late September 2026.”
Transitioning From Cryptocurrency Operations
As AI infrastructure becomes operational, Hyperscale Data intends to systematically reallocate electrical capacity at the Michigan location currently dedicated to Bitcoin mining activities. The facility presently operates approximately 28 megawatts supporting Bitcoin mining.
Bitcoin mining will persist at the company’s Montana operation. Limited mining activity may continue at Michigan throughout the transition phase.
Hyperscale Data’s current holdings include 726.9 bitcoin, representing approximately $45.9 million in value. The company maintains an active $300 million at-the-market equity offering program facilitated through Spartan Capital Securities.
Financial Scope and Implications
The magnitude of this MSA is particularly noteworthy considering Hyperscale Data’s present financial profile. The organization maintains a market capitalization of merely $124 million and recorded $121 million in revenue during the trailing twelve months.
The company also bears a $113 million debt burden, and InvestingPro analysts have highlighted ongoing cash consumption ā a consideration that may impact its capacity to fund the $100-120 million infrastructure development.
Executive Chairman Milton “Todd” Ault III described the agreement as “a significant milestone,” stating the company’s long-term objective involves developing more than 300 megawatts of aggregate power capacity at the Michigan campus.
Company leadership emphasized that any expansion beyond the initial deployment phase remains conditional and depends on regulatory clearances, securing financing, infrastructure readiness, and finalizing utility partnerships.
Most recently, Hyperscale Data disclosed the acquisition of 8 bitcoin through open market purchases during the week concluding June 21.


