THE CONTROL LAYER BETWEEN GRID AND AI COMPUTE

Active power orchestration for grid-compliant ORv3 / ORW AI racks.

The grid won't bend for AI. So we make the rack bend for the grid.

DPO responds to AI power transients and coordinates BBU + PSU output to meet ERCOT / NERC ride-through and post-fault recovery requirements at rack level.

Book a technical call Read the architecture

THE PROBLEM

Grid compliance just became a continuous control problem.

In July 2024, six data centers in Texas tripped offline simultaneously when the grid stumbled. AI training and inference create power transients on millisecond time scales — the kind of behavior grid operators can no longer ignore.

Eight jurisdictions are now writing rules so that does not repeat. Beginning January 2028, ERCOT NOGRR 282 makes voltage ride-through and post-fault active power recovery mandatory for large loads — including data centers above the size threshold.

Read the compliance landscape →

WHAT WE BUILD

Two coordinated layers. One closed loop.

Rack-level control layer

DPO Engine

Forecasts GPU load ahead of the transient. Computes setpoints for the rack power shelf. Runs as a software process alongside the workload scheduler.

Interfaces: Redfish · PMBus · OCP ORv3 PMI

Learn more →

Rack-mounted hardware module

DPO Gateway

Pin-compatible replacement for the PMI module. Coordinates BBU discharge and PSU output shaping in real time. Designed against ORv3 PMI Specification 1.0.

Form factor: OCP ORv3 PMI-compatible · Cert target: UL + OCP 2027 Q4

View spec →

ARCHITECTURE

Predict, orchestrate, comply.

DPO sits between the data center EMS and the ORv3 rack — predicting the transient, computing the setpoint, and enforcing the envelope at the shelf. Inline with ORv3 PMI specification, brand-neutral when contributed to the OCP open spec.

DATA CENTER EMSFleet scheduler · workload orchestratorPOWER GRIDDemand · capacity · frequency schedulerDPO ENGINE — CONTROL PLANEML-driven rack-level power predictionVRT POLICYVoltage ride-through envelope enforcementPFAPR RECOVERY CONTROLPost-fault active power recovery slopeCONTROL PLANEDPO GATEWAY — SHELF PMI EXTENDRack-mounted power orchestration moduleVRT EDGE POLICYLocal VRT enforcementBBU/PSU ORCHESTRATORDischarge + output shapingRACK POWER PREDICTIONSub-ms transient forecastTELEMETRY AGGREGATORFleet data collectionAC INPUTGrid powerORv3 RACKACPOWER SHELFAC → DC conversionDPO48V DC48V DC BUSBBU SHELFStandard backupCharge ⇄ dischargeDPOBBU — VRT/FRT/PFAPRGrid compliance eventsCharge ⇄ dischargeDPOAI SERVER ×6+GPU compute load
DPO Gateway Specification — XMight's open contribution to OCP
See architecture detail →

TRACTION

Where we are.

XMight is a member of the OCP Startup Program 2026 and is currently in pilot deployment with a North American Edge AI data center operator.

FAQ

Common questions about XMight DPO.

What is XMight DPO?
DPO (Dynamic Power Orchestrator) is a two-layer system — Cloud-side DPO Engine for prediction + rack-mounted DPO Gateway for execution — that delivers grid-compliant power orchestration for OCP ORv3 / ORW AI racks. It coordinates BBU discharge and PSU output in real time to pass NOGRR 282 ride-through without throttling GPU compute.
What is NOGRR 282?
NOGRR 282 is the ERCOT Nodal Operating Guide Revision Request that mandates voltage ride-through (VRT) and post-fault active power recovery (PFAPR) for large loads — including AI data centers above a size threshold. Enforcement begins January 2028. SPP and other North American grid operators are adopting similar codes.
Why can't ORv3 / ORW racks pass NOGRR 282 without help?
OCP ORv3 / ORW racks distribute battery backup (BBU) across each rack for energy efficiency but provide no orchestration layer. On grid restoration, PSU startup is hard-coded with randomized delay — taking seconds to minutes. This is physically incompatible with the sub-second PFAPR recovery NOGRR 282 requires.
How is DPO different from a traditional UPS?
Traditional UPS rebuilds operate at facility level — months to over a year of mainline electrical construction. DPO operates at rack level. The DPO Gateway is a pin-compatible OCP PMI module; install is rack-level hot-swap, measured in minutes per rack. IT and OT are fully decoupled.
Does DPO throttle GPU compute?
No. DPO orchestrates the power layer (BBU + PSU) without touching the compute layer. AI workloads continue at full throughput during grid events — in contrast to software GPU throttling approaches, which interrupt synchronous training with straggler effects and NCCL communication failures.
What is POM, and how does it relate to DPO?
POM (Power Orchestrator Module) is the open specification XMight contributes to the OCP Rack & Power working group, defining a standard interface for rack-level power orchestration. DPO is XMight's commercial implementation of POM — the DPO Gateway hardware conforms to the POM spec; the DPO Engine cloud layer adds XMight's proprietary prediction and policy logic on top.
Is XMight a member of the Open Compute Project (OCP)?
Yes. XMight Smart Energy Corp. is a member of the OCP Startup Program 2026 — one of a small global cohort and the sole Asia representative in the current cohort. We contribute to the OCP Rack & Power working group, advancing the POM (Power Orchestrator Module) specification as an open standard.
When does NOGRR 282 enforcement begin?
January 2028. From that date, new data centers must demonstrate NOGRR 282 compliance before they are allowed to interconnect to the ERCOT grid. Existing data centers that cannot comply will be first in line to be disconnected when the grid is stressed. SPP and other operators are aligning to similar 2028 timelines.

Talk to the team before 2028 becomes this year.

Technical inquiry Read the architecture