No equipment control
Headroom is not equipment control, charger management, battery dispatch, DERMS, or VPP software.
Headroom
Headroom helps counterparties understand when flexible compute was supportable, when it needed to stand down, what obligation outranked it, and what telemetry supported the decision.
Headroom is the black-box evidence layer for fragmented sites. It records capacity states, constraints, and replayable evidence; it does not operate the site.
What Headroom does
Headroom reads telemetry and historical data from mixed-energy sites, scores telemetry quality, and applies ranked obligations such as EV charging, battery reserve, utility events, thermal limits, and site-priority loads.
It produces conservative capacity states, emits reason codes and replayable evidence records, and supports audit, settlement support, dispute reconstruction, pilot screening, and buyer diligence.
Headroom has been exercised through public-data replay across multiple energy and EV data classes. Synthetic Headroom overlays are separately labeled. The replay corpus is not a production deployment, does not control equipment, and does not imply third-party endorsement.
What Headroom is not
Headroom is not equipment control, charger management, battery dispatch, DERMS, or VPP software.
Headroom is not a GPU scheduler, workload scheduler, or compute marketplace.
Headroom is not legal adjudication, third-party certification, or a production customer deployment.
Why this matters
Power-backed compute sites can involve a site owner, charger operator, battery owner, utility or demand-response aggregator, compute operator, and workload or customer. Each party can have different obligations, data access, and economic exposure.
Headroom is designed for non-vertically-integrated sites where these parties need a neutral record of what was supportable, what was constrained, and what data was available at the time.
Public replay corpus
Headroom has been exercised through public-data replay across commercial building-plus-EV profiles, EV charging-session data, EV demand-response data, urban charging availability/demand data, high-frequency electrical load telemetry, and DOE EV charging-session data. Synthetic Headroom overlays are separately labeled. These replay results demonstrate read-only normalization, telemetry-quality degradation, ranked-obligation classification, reason-coded evidence records, and audit/replay behavior. Public-data replay is not a production deployment, does not control equipment, and does not imply third-party endorsement.
| Source class | Replay use | Public-use caution |
|---|---|---|
| NREL/OEDI: commercial building + EV charging profiles | Commercial building load plus EV charging load replay into capacity-envelope classification. | Public profiles are replay inputs, not a customer deployment or third-party endorsement. |
| ACN-Data: real EV charging sessions | EV charging obligation replay with redacted public session context and synthetic compute demand. | Do not display raw session, station, or space identifiers in public excerpts. |
| Jeju EV DR: EV charging + demand-response/event context | EV load replay with public external demand-response context as a superior-obligation signal. | Attribute the public dataset and do not imply author, Figshare, or third-party endorsement. |
| UrbanEV: urban charging availability/demand | Urban EV demand and availability replay into capacity evidence with redacted zone labels. | Do not expose station or node identifiers in public excerpts. |
| REFIT: high-frequency electrical load telemetry | Telemetry-quality degradation replay for clean, partial, and stale load telemetry conditions. | Avoid household identity and appliance-level public excerpts. |
| DOE EV Data Collection: DOE EV charging-session data | Read-only EV charging-session replay with public session data separated from synthetic overlays. | Do not display raw vehicle, charger, fleet, facility, or session identifiers. |
| Fixture adapters: separately labeled fixtures | Public-data-ready adapter testing, normalization, telemetry-quality degradation, and reason-coded demo outputs. | Fixtures are not actual public replay records and should not be described as public-dataset backtesting. |
Black-box recorder
Headroom is not the system that runs the site. It is the black-box evidence layer that records what capacity was supportable, what constraint blocked it, and why a lower-priority load should be admitted, limited, preserved, denied, or redirected by a downstream system.
Modular compute context
Modular compute near power-rich EV charging or battery-backed sites is an emerging category. Headroom does not claim Tesla integration, Supercharger integration, Megapod integration, or affiliation with any vehicle, charging, battery, or compute platform.
Headroom is for fragmented, non-vertically-integrated sites where parties need neutral evidence before they contract, operate, reconcile, or dispute flexible compute capacity.
Source and caution
Public-data replay uses third-party public datasets and separately labeled synthetic Headroom overlays. Source-specific attribution and license details are maintained in the Headroom source registry.
No third-party endorsement is implied. Dataset owners, authors, DOE, NREL, NLR, Alliance, ACN, Caltech, PowerFlex, Figshare, UrbanEV authors, Zenodo, REFIT authors, Livewire, Nature, Dryad, GitHub, and other third parties do not endorse, review, approve, validate, certify, or fund Sinteag or Headroom.
Public excerpts should not expose raw dataset identifiers, station IDs, charger IDs, vehicle IDs, household IDs, customer IDs, session IDs, or facility IDs.
Next step
For qualified sites, Sinteag can discuss a read-only Headroom pilot, review public replay evidence, or explore mixed-energy site screening.