Hyperscale Data, Inc. (NYSE American: GPUS) has engaged commercial real estate advisory firm Newmark Group to identify lenders for long-term debt financing for its Michigan AI data center campus. The proposed financing is intended to fund construction and expansion while reducing the company’s reliance on issuing additional shares. No loan has been secured, and the financing amount and terms have not been disclosed.
Newmark to Explore Institutional Financing Options
The engagement gives Hyperscale Data access to Newmark’s data center and digital infrastructure capital markets team. Newmark will identify potential lenders and help evaluate financing structures, repayment schedules and other terms suited to the Michigan project’s development timeline.
The distinction between arranging financing and securing a loan is important. Hyperscale Data has not announced a financing commitment, target borrowing amount, interest rate or maturity. Any transaction will depend on lender underwriting, due diligence, definitive agreements and closing conditions. The company has also acknowledged that acceptable terms may not be available and that further equity financing could still be necessary. Nasdaq+1
Newmark brings experience with large-scale data center transactions. In May 2025, it announced that it had arranged a $7.1 billion construction loan for a major AI data center development in Abilene, Texas. In June 2026, it reported arranging $975 million in financing for a newly constructed data center in Northern Virginia. These transactions demonstrate experience in the sector, although they do not establish what terms Hyperscale Data might obtain. Newmark+1
A Customer Agreement Provides a Basis for Development
Hyperscale Data’s Michigan campus is operated by Alliance Cloud Services, an indirect wholly owned subsidiary. According to the company, ACS has signed a master services agreement with a California-based neocloud provider for colocation and related data center services supporting an initial 20 megawatts (MW) of critical AI computing capacity.
The agreement has an initial 10-year term and includes two five-year extension options exercisable by the customer. It also grants the customer a right of first offer on an additional 32 MW, if available, potentially bringing the total deployment to 52 MW. The additional capacity is not guaranteed and remains subject to contractual terms and development requirements. Nasdaq+1
Hyperscale Data expects the first 10 MW to be energized in December 2026, with revenue-generating activity expected to begin around that milestone. The schedule remains dependent on completing infrastructure and commissioning work.
That timeline is central to the financing effort. A data center moving from construction into commercial operation has different funding needs and risk considerations at each stage. Lenders will need to assess project costs, construction progress, the expected timing of customer deployments and the reliability of projected cash flows. The existing customer agreement provides a commercial framework, but it does not eliminate execution risks.
AI Infrastructure Demand Raises the Stakes for Capital
The financing effort comes as data center developers face growing demand for computing capacity alongside significant infrastructure requirements. AI facilities need not only servers and networking equipment but also substantial electrical capacity, cooling systems and supporting infrastructure. These costs can make financing strategy a major factor in how quickly a project can be developed.
The International Energy Agency’s 2025 report, Energy and AI, projected that global data center electricity consumption would more than double to approximately 945 terawatt-hours annually by 2030, with AI a major driver. The IEA’s later analysis continues to identify power supply, equipment availability and other infrastructure constraints as factors shaping expansion. IEA+1
For data center operators, rising demand does not automatically translate into profitable capacity. Projects must secure financing, complete construction, obtain sufficient power and bring computing infrastructure online before contracted services can generate revenue.
Debt financing can help companies fund these investments without immediately issuing more shares. However, borrowing creates fixed obligations, including interest and principal repayments, and may involve collateral requirements and restrictive covenants. Equity issuance can dilute existing shareholders, while debt can increase financial pressure if construction is delayed or revenue arrives later than expected.
Hyperscale Data says it will evaluate prospective financing based on interest expense, repayment obligations, collateral, covenants and other terms. The final structure will therefore matter as much as the amount raised.
Financing Remains the Key Unresolved Milestone
Executive Chairman Milton “Todd” Ault III and CEO William Horne said the company wants a long-term institutional financing structure that limits dilution while supporting construction and customer deployments. Those statements describe management’s objectives, not a completed financing outcome.
The Michigan project now faces two important milestones: completing the infrastructure needed to energize the first 10 MW and determining whether lenders will provide capital on acceptable terms. The company has not disclosed a timetable for completing the financing process.
For investors and the wider AI infrastructure market, Newmark’s appointment is a step toward funding the campus, rather than confirmation that the capital is available. The project’s progress will depend on lender interest, borrowing costs, construction execution and the pace at which contracted capacity becomes operational.
Hyperscale Data’s approach reflects a broader challenge in AI infrastructure finance: matching long-term capital commitments with projects whose costs arrive before their expected revenue. Whether debt can reduce the company’s reliance on equity will become clearer only after financing terms are agreed and the project advances toward commercial operation.
Market Landscape
AI data center development is becoming a capital-intensive segment of digital infrastructure. Operators must finance land, power infrastructure, cooling, networking and computing equipment before facilities can generate revenue.
The International Energy Agency projected in its 2025 Energy and AI report that global data center electricity consumption could reach approximately 945 TWh annually by 2030, more than double its then-current level. IEA
Hyperscale Data’s financing initiative highlights three industry dynamics:
- Institutional capital: Specialized financing advisers can help developers approach lenders familiar with large-scale data center projects.
- Equity dilution versus debt obligations: Debt may reduce immediate share dilution but introduces interest costs, repayment requirements and potential collateral commitments.
- Customer-backed development: Long-term service agreements can help establish a commercial basis for infrastructure investments, although construction, commissioning and customer expansion remain uncertain.
The critical question is whether Hyperscale Data can secure financing on sustainable terms while meeting its planned December 2026 energization milestone.
Top Insights
- Hyperscale Data hired Newmark to identify potential lenders for long-term debt financing of its Michigan AI data center campus.
- The Michigan campus has a customer agreement covering an initial 20 MW, with a conditional pathway to a potential 52 MW deployment.
- The company expects its first 10 MW of critical AI computing capacity to be energized in December 2026, subject to construction and commissioning.
- Newmark’s previous data center financing assignments provide relevant experience, but they do not guarantee financing for Hyperscale Data.
- No loan has been secured, and the company may still need additional equity depending on financing terms and project requirements.
Get in touch with our fintech expert

