Exascale Labs is converting an estimated $300 million customer pipeline into binding and non-binding revenue agreements across GPU-as-a-Service and AI infrastructure, signaling that demand for modular data center capacity remains strong despite supply constraints.
Exascale Labs, a San Francisco-based AI infrastructure provider merging with D. Boral ARC Acquisition I Corp. (Nasdaq: BCAR), said it signed a non-binding memorandum of understanding with AI services company RUTILEA for the potential development and commercialization of as much as 20 megawatts of data center capacity. The company also disclosed a $71.4 million three-year compute services agreement with Dimension AI and a $53 million non-binding MOU with AI Nova for GPU servers and technical services.
"These agreements and non-binding MOUs represent tangible progress in advancing our estimated $300 million customer opportunity pipeline into both near-term and multi-year recurring revenue," Chief Executive Officer Hoansoo Lee said.
The RUTILEA MOU targets 20MW of capacity — enough to power roughly 17,000 U.S. homes — using Exascale's modular data center designs, which include high-density cooling, HVDC power and energy storage solutions. The company's asset-light model sources GPU compute from third-party data centers globally, then resells reserved and on-demand access alongside cluster management services. Exascale's platform is built for large language model training, fine-tuning and high-concurrency inference workloads.
The Infrastructure Gap Exascale Is Trying to Fill
The deals come as hyperscalers and AI startups compete for limited GPU capacity, with Nvidia's H100 and B200 chips facing allocation lead times that stretch into months. Exascale's approach — offering modular infrastructure that can be deployed faster than traditional brick-and-mortar data centers — positions it as an alternative for customers who cannot secure capacity from AWS, Microsoft Azure or Google Cloud. The company recently showcased its modular data center, HVDC and solid-state transformer offerings alongside partner Compal Electronics at COMPUTEX Taipei 2026.
The $71.4 million Dimension AI agreement locks in dedicated GPU compute capacity over three years, providing Exascale with a multi-year revenue stream in its GPU-as-a-Service business. The $53 million AI Nova MOU contemplates initial orders for GPU servers and related technical services, though both MOUs remain non-binding and subject to definitive agreements.
What It Means for Investors
Exascale is completing a business combination with D. Boral ARC Acquisition I Corp. under an agreement signed Jan. 11, 2026. The SPAC merger gives Exascale access to public markets to fund its infrastructure buildout at a time when AI compute demand continues to outstrip supply. If the RUTILEA, AI Nova and Dimension AI agreements convert into definitive contracts, Exascale would have secured at least $124.4 million in committed or anticipated revenue against its $300 million pipeline — roughly 41% of the total opportunity.
The broader AI infrastructure market, valued at more than $50 billion annually by industry estimates, remains dominated by hyperscalers. Exascale's modular approach targets a niche: customers who need dedicated capacity faster than traditional data center leases allow. The company's ability to convert non-binding MOUs into signed contracts will determine whether its pipeline translates into actual revenue.
This article is for informational purposes only and does not constitute investment advice.
