Remove design errors before construction starts,
and save the cost of rework
and save the cost of rework
KONST uses NVIDIA Omniverse to turn design drawings into a one-to-one three-dimensional model. Before construction starts, we verify space, heat flow, hydraulics, power redundancy, and network architecture, and deliver a design that construction can be based on, together with a list of problems resolved before construction. The model is the single source of truth, and every design change is verified against the same model.
Errors in a high-density data center only show up after power-on
Traditional data center design experience mostly comes from low power density loads. When it is applied to GPUs at tens of kilowatts per rack, hot spots, airflow short-circuiting, and power distribution problems often appear only after power-on, and fixing them costs far more than adjusting before construction.
Design, delivery, and operations as one complete process
The model is not a one-time acceptance document. Before construction it is used to verify the design, at handover it is calibrated to match the site, and after go-live it is connected to operating data and keeps answering capacity questions.
- Problem
Design complete, not yet awarded to a contractor
Verification has the widest coverage, because space, heat flow, hydraulics, and network can all still be changed on the drawings.
ProblemContract awarded, equipment not yet on site
Rerun the simulation with the equipment models actually ordered, and correct the rack layout and piping before the equipment arrives on site.
ProblemExisting data center retrofit
Load capacity, floor height, incoming power capacity, and existing on-site piping are all fixed conditions. Running the model first shows the upper limit of what the retrofit can do.
SolutionDesign verification
At the drawing stage, we resolve spatial clashes, hot spots, hydraulic imbalance, and power redundancy problems, and deliver verification results that construction can be based on.
- Problem
Equipment on site, not yet powered on
Compare the model with the actual layout, find the differences between simulation and site, and resolve them before power-on.
SolutionHandover calibration
After power-on, the model's behavior is tuned to match the actual data center, and from then on the model is the reference baseline for the site.
- SolutionExtension
Operations optimization
The model is connected to operating data and updated continuously, so capacity and layout decisions are based on actual conditions rather than design assumptions.
Design verification areas and main outputs
Space
What is verifiedPhysical clashes between piping, power paths, cable trays, and structure
Main outputClash detection report
Heat flow
What is verifiedCFD analysis of airflow distribution and cooling efficiency at full load
Main outputHeat flow simulation results and redundancy level
Hydraulics
What is verifiedLiquid cooling pipe pressure drop, flow velocity, and branch flow distribution
Main outputHydraulic balance simulation results
Failure drills
What is verifiedTemperature rise and response window after a CDU or air conditioner shuts down
Main outputBasis for judging whether the backup is sufficient
Power
What is verifiedCheck of the N+1 or 2N power distribution redundancy level
Main outputArchitecture review comments
Network
What is verifiedLeaf-Spine topology, non-blocking switching, and oversubscription ratio
Main outputLogical topology diagram and port connection matrix
Cabling
What is verifiedCable path planning in a three-dimensional environment
Main outputList of the actual length of each cable
After go-live,
the model changes from a verification tool to an operations tool
Once the model is updated in sync with the actual data center, the digital twin can be calibrated with digital information in real time. Later expansion planning, asset inventory, and failure assessment can then all be done on this same calibrated model.
Shorten the time from equipment to usable compute
The model stays in sync with telemetry, assets, and existing tools, so it is always current and verification before deployment does not require rebuilding the model.
Find gaps and errors in asset data
Once data is consolidated into one model, fields that are unmaintained, duplicated, or inconsistent with the site show up immediately, and the accuracy of asset records improves.
Consolidate data spread across systems
Data from MEP monitoring, power management, and asset management is brought onto one platform for analysis, and cross-department changes and reviews happen on the same model.
FAQ
We already have design drawings. Do we still need digital twin verification?
What data do we need to provide to use a digital twin?
If the design changes midway, does the verification need to be rerun?
Our data center is already running. Can we still do a digital twin?
Services often evaluated together
Want to verify the design before construction starts?
Send us your design drawings and equipment list, and we will reply with the verification scope and schedule.