Solmar Insights
Duos Technologies Group has formalized five-year hosting agreements with Axe Compute Inc., securing a total of 55 MW of dedicated AI data center capacity across several US sites. The contracts represent over $500 million in base payments, excluding usage-based charges, with commencement of operations projected for late 2026 into early 2027.
Key figures
55 MW AI data center agreements
Five-year contract term
More than $500 million in base payments
Structure of the hosting agreements
The five-year hosting agreements executed between Duos Technologies Group and Axe Compute Inc. reserve a total of 55 MW of AI compute capacity. These contracts span multiple purpose-built data center sites in the United States, tailored to support next-generation GPU systems critical for artificial intelligence workloads. The capacity is contracted on a long-term basis, with billing tied to successful completion of site readiness, performance testing, and acceptance by Axe Compute. Importantly, the more than $500 million value attributed to these agreements covers only the contractual base payments, not accounting for additional usage-based electricity or services charges.
Agreements of this scale underscore the growing trend toward long-term power commitments and colocation in the data center industry, particularly as demand for AI and high-performance computing continues to surge. The inclusion of renewal options and expansion rights creates a strategic foundation for future scale, with Duos retaining majority project ownership and Axe Compute evaluating minority investments under non-binding term sheets.
The transaction further reflects the operational complexity of advanced AI facilities, with design mandates for ultra-high density, cooling, and uptime. Both companies have positioned the agreements not just as a revenue driver, but as a step toward broader infrastructure partnerships aiming to deploy and monetize large-scale, enterprise-grade GPU compute environments.
Timeline and deployment details
Readiness for initial capacity under these agreements is forecast to begin in late 2026, with ongoing deployments continuing into early 2027. All deployments are subject to Axe Compute’s acceptance following successful construction, commissioning, and performance testing. The multi-site development strategy offers flexibility for geographic placement based on power availability, interconnection, and proximity to demand centers.
This staggered project readiness aligns with the broader US data center construction cycle, where major deployments may take several years due to permitting, grid interconnection, power procurement, and equipment lead times. The dynamics of GPU availability and liquid cooling technology are materially reshaping development timelines, necessitating detailed engineering and substantial upfront investment. Delays at any stage of construction or commissioning could impact when revenue from these agreements starts to be recognized.
For both Duos Technologies and Axe Compute, the structure ensures that substantial capital is only deployed in lockstep with customer commitments, reflecting industry risk management best practices for hyperscale and AI-oriented infrastructure.
Financing and investment terms
While the contracts themselves lay out the commercial arrangements and baseline payments, the parties have also entered non-binding term sheets addressing possible minority investments by Axe Compute in the entities responsible for project execution. Any such equity participation would be subject to further negotiation, definitive agreements, and required approvals, while Duos is anticipated to keep majority control for governance and operational direction.
This capital structure approach, potential minority customer investment, mirrors evolving data center financing trends, where anchor customers sometimes seek direct stakes to secure capacity while aligning interests with the developer. Such structures can support competitive pricing for users and enhance access to project financing by demonstrating robust revenue commitments over a multi-year horizon.
The financing mechanism and contractual strategy are both crucial in a market contending with high upfront capital requirements, uncertain power procurement, and rising development costs, especially for facilities capable of supporting dense GPU clusters for AI training and inference.
AI data center requirements
The AI hosting agreements were structured around next-generation GPU requirements, which impose far greater demands on power density, cooling, and electrical redundancy than traditional data center deployments. These design challenges have become a central focus for both operators and enterprise customers looking to deploy large-scale AI and machine learning workloads.
Higher power draw per rack and the need for advanced cooling solutions, often including liquid cooling, drive both capital and operational expenditure. Designing for high availability and modular expansion enables tenants to bring capacity online incrementally while keeping infrastructure scalable as AI compute requirements evolve. These considerations shape site selection, grid interconnection planning, and the overall engineering workflow for data center operators like Duos Technologies.
For Axe Compute, access to reserved, enterprise-grade GPU capacity is critical for maintaining service levels as client AI workloads increase. These hosting deals demonstrate how specific data center technical requirements increasingly factor into both commercial agreements and project execution in the AI infrastructure sector.
Market significance for AI and infrastructure
This agreement represents one of the larger known commitments for enterprise AI hosting capacity by a specialist infrastructure platform in 2026. The scale, long-term nature, and technical specificity highlight a maturing commercial model for U.S. data center growth driven by generative AI workloads and enterprise GPU demand.
The multi-site structure provides flexibility to respond to both regional grid conditions and shifting locus of compute demand, potentially benefiting from tax incentives or regional power pricing arbitrage. For institutional investors and developers, these transaction structures reinforce the rising need for bespoke data center infrastructure tied to contractual offtake from AI service providers.
As other cloud and hyperscale buyers seek similar arrangements, recurring revenue visibility and non-cancellable terms in this deal provide a model for de-risking project capital outlays. The shared signals from Duos Technologies and Axe Compute point to intensifying competition for construction-ready sites and power allocations capable of supporting next-generation AI clusters at scale.
What this means for buyers
For energy and digital infrastructure buyers, this deal signals mounting demand for high-density, power-intensive data center projects to serve AI customers. Committed long-term offtake, significant capital deployment, and investment structures with customer participation are setting new standards for project viability and risk-sharing. Buyers should track execution risk around power procurement, interconnection, and commissioning milestones as these become critical drivers of both revenue timing and asset value in the AI-hosting sector.
