Specifications
Product Type
Coolant Distribution Unit (CDU)
Flagship Model
ROL2300 — high-power AI workloads
Mid-Density Model
ROL1100 — modular, AI-ready in-row
Target Loads
AI servers, GPUs and TPUs
Integration
Boyd cold plates, manifolds, immersion-ready systems
Monitoring
Remote diagnostics and smart monitoring tools
Cold Plate Construction
Precision CNC machining + high-quality brazing
Leak Testing
100% leak tested in-line
Thermal/Flow Testing
100% in-line thermal and flow performance testing
Environmental Testing
Temperature cycling, power cycling, thermal shock
Logistics Testing
Packaging and shipping tests in-house
Mounting
Aluminum mounting frame for high mounting force
Loop Design
Multiple cold plates joined to manifolds via quick disconnects
Production Volume
5,000,000 liquid cold plates delivered to hyperscalers
Ecosystem
Eaton Boyd Thermal Chip to Ambient thermal management
Overview
Boyd (now operating within Eaton Boyd Thermal) builds liquid cooling from the chip outward. Its coolant distribution units split into the ROL2300 for high-power AI workloads and the ROL1100, a modular in-row unit aimed at mid-density AI racks where over-provisioning cooling capacity wastes both capex and floor space. The ROL1100 is designed around smaller rack footprints and integrates with Boyd cold plates, manifolds and immersion-ready systems.
The differentiator is vertical integration. Boyd designs and manufactures liquid cold plates in volume — five million delivered to hyperscalers to date — using precision CNC machining and high-quality brazing, with complex internal surface area engineered to maximise cooling capacity for GPUs and CPUs. Each plate is 100% leak tested and 100% thermally and flow tested in line.
Multiple cold plates connect to manifolds with quick disconnects to form a loop that directs coolant across every processor on a board, held in place by an aluminium mounting frame that enables high mounting force. In-house testing covers temperature cycling, power cycling, thermal shock, packaging and shipping.
Key Benefits
Purpose-built CDUs for AI density — ROL2300 for high-power racks, ROL1100 for mid-density where right-sizing matters. Cold plate to CDU from one supplier removes interface risk between chip-level and row-level cooling. 100% leak tested and 100% flow/thermal tested before shipment. Hyperscale manufacturing pedigree — five million liquid cold plates delivered. Remote diagnostics and smart monitoring simplify upkeep and protect uptime.
Applications
Direct-to-chip cooling of AI GPU and TPU racks, mid-density AI server deployments, immersion-ready retrofits, and cold-plate supply for board-level loops connecting multiple accelerators through manifolds and quick disconnects.
Request a Quote — BOYD ROL2300 CDU — LIQUID-COOLED COOLANT DISTRIBUTION UNIT FOR AI WORKLOADS
QS Compute — global B2B supply of AI computing hardware, edge AI systems and accelerators. Volume pricing, 15-day sample lead time.
Get Your Quote →