PHASURGE · Product system
A practical product portfolio built from phase-change cooling, direct-cool hardware and AI control.
Server two-phase direct cooling
Prototype · Design targets
Built for high heat-flux AI compute: 60kW per rack, PUE ≤ 1.05.
BESS direct cooling / heating
R&D validation
5kW wide-range -30~50°C, COP 4.5, for C&I and utility-scale BESS scenarios.
EV direct-cool / direct-heat test rig
Product definition · Application-specific configuration
Dual-mode cooling & heating, sub-second mode switching for a standardised testing workflow.
The DC line spans rack CDUs, cabinet CDUs and two-phase cold-plate modules for staged adoption in AI compute facilities.
P2-W200C
Water-to-refrigerant cabinet 200kW CDU
P2-W60R
Water-to-refrigerant rack 60kW CDU
P2-A60C
Air-to-refrigerant waterless cabinet 60kW CDU
P2-A20R
Air-to-refrigerant waterless rack 20kW CDU
P2Edge-6U
All-in-one edge unit with plug-and-run deployment
GPU cold plate
Two-phase cold plate / module
CPU cold plate
Low-resistance uniform-temperature cold plate
HBM/HPM cooling plate
Cooling plate for optical modules and high-bandwidth memory
Multi-node cold-plate module
Parallel multi-node cooling module
Start with electricity costs, then explore dispatch and equipment conditions.
Plan electricity, solar and storage dispatch, and compare the cost of operating strategies.
Review battery conditions and thermal alerts to plan equipment maintenance.
Choose a scenario to compare cooling energy and electricity costs.
Operating conditions: 1,000 kW · 8,000 hours/year · 0.8 CNY/kWh
Annual electricity savings
¥576,000Annual energy savings: 720,000 kWhCooling energy reduction
30%90 kWElectricity-savings payback
2.1yearsSelect equipment to view electrical power
Electricity per million Tokens
Compute and cooling electricity only; excludes other facilities, hardware depreciation and model-service fees.
Enter average electrical power, not heat-removal capacity. Costs are in CNY. Energy = power × time; electricity cost = energy × tariff.
PHASURGE · Engineering
