Hyperscale Data Accelerates Michigan Campus Into Combined AI Data Center and Robotics Hub Following AGIBOT Partnership
AI / Machine Learning | 4 min read
Hyperscale Data, Inc. (NYSE American: GPUS), an AI data center company anchored by Bitcoin, has announced it is accelerating the enhancement of its Michigan operations into a combined AI data center and robotics hub — following its recently executed agreement with AGIBOT PTE. LTD., a developer of intelligent robotics technology. Through its subsidiary Omnipresent Robotics, LLC, Hyperscale Data will serve as AGIBOT's sole North American partner, deploying AGIBOT G2 robots for teleoperation, data collection, and pilot deployments. The company is advancing the reconfiguration of key sections of its existing 617,000 square foot facility on its 34.5-acre campus in Dowagiac, Michigan — with plans to expand to 83 acres — to support AI infrastructure, robotics deployment, and large-scale data generation. The site currently operates approximately 30 megawatts (MW) of power capacity, with potential to expand to over 300 MW over time. Hyperscale Data expects to hire more than 500 employees over the next three years, with roles spanning robotics engineering, AI data operations, infrastructure, and operations staff.
"We are enhancing our Michigan operations into a scaled AI and robotics platform that supports both high-performance compute and real-world data generation. This next phase of development positions us to support evolving AI workloads while creating high-quality jobs in the region."
— William B. Horne, Chief Executive Officer, Hyperscale Data
The Physical AI Data Thesis — Real-World Data for the Next Phase of AI
Hyperscale Data is initially dedicating more than 100,000 square feet within its existing facility to AI and robotics operations. Within this footprint, the company plans to develop robotics assembly and testing capabilities, build real-world environments for data collection and system validation, and integrate compute infrastructure with robotics-driven data generation. The Michigan campus is being designed to support three core data types: machine-generated data from robotics operating in real environments, human (egocentric) data capture for contextual learning, and training workflows for advanced AI models. This data — including information from real failures, torque and friction measurements, and repeated task attempts — is expected to be commercialised and sold to frontier AI developers in US markets, supporting domestic AI development and deployment. Omnipresent Robotics intends to begin teleoperation, data collection, and pilot deployments using AGIBOT G2 robots, positioning Hyperscale Data as a complete domestic loop — supply chain, computing, training, evaluation, assembly, and deployment — for embodied AI infrastructure.
"This initiative builds on our existing infrastructure and expands it into a more comprehensive AI ecosystem."
— Milton "Todd" Ault III, Executive Chairman, Hyperscale Data
Key Takeaways
- • Hyperscale Data (NYSE American: GPUS; CEO William B. Horne; Executive Chairman Milton "Todd" Ault III; AI data center company anchored by Bitcoin) has announced it is accelerating its Michigan operations into a combined AI data center and robotics hub following its agreement with AGIBOT PTE. LTD. (intelligent robotics developer). Through subsidiary Omnipresent Robotics LLC, Hyperscale Data will serve as AGIBOT's sole North American partner — deploying AGIBOT G2 robots for teleoperation, data collection, and pilot deployments at its Dowagiac, Michigan campus.
- • Facility and infrastructure: 34.5-acre campus at 415 E. Prairie Ronde St., Dowagiac, Michigan (expanding to 83 acres); 617,000 sq ft total facility; 100,000+ sq ft initially dedicated to AI and robotics operations. Current power capacity: ~30 MW; potential expansion to 300+ MW over time. The facility is being reconfigured for robotics assembly, testing, and deployment alongside AI model training and validation — integrating compute infrastructure with robotics-driven data generation.
- • Three data types being generated: machine-generated data from robotics operating in real environments; human (egocentric) data capture for contextual learning; training workflows for advanced AI models. Data sources include real failures, torque and friction measurements, and repeated task attempts — representing the high-quality physical AI training data that traditional digital datasets cannot provide. Data will be commercialised and sold exclusively to US markets, supporting domestic AI development and deployment.
- • Target customers and hiring: frontier AI developers — those advancing large-scale models and next-generation robotics systems — and high-performance computing platforms. 500+ employees to be hired over three years: robotics engineers, AI data specialists, infrastructure personnel, and operations staff. Hyperscale Data's thesis: the next phase of AI will increasingly depend on real-world data and physical system training in addition to traditional digital datasets, creating a structural shortage in high-quality physical AI training data that its Michigan hub is designed to address.
- • Company context: Hyperscale Data's Michigan campus has been a focal point of its AI infrastructure strategy — previously including NVIDIA Blackwell GPU systems build-out (September 2025) and a $180–200M 2026 revenue guidance issued March 11 2026. The AGIBOT partnership and robotics hub announcement represents a strategic pivot from pure digital infrastructure (Bitcoin mining, traditional hosting, GPU cloud services) into embodied AI and physical AI data generation. Hyperscale Data's ambition is a complete domestic loop: supply chain, computing, training, evaluation, assembly, and deployment — without dependence on external parties — for US-based embodied AI development.