Virtual power plant VPP

What it is (precise)

The definition the US Department of Energy uses, quoted by Idaho National Laboratory in its January 2026 report on VPP architecture and resilient design, is that VPPs are "aggregations of DERs that can balance electrical loads and provide utility-scale and utility-grade grid services like a traditional power plant".

INL adds the operational content: a VPP uses a digital platform to forecast, optimize and dispatch thousands of devices so they operate as a unified portfolio rather than as isolated assets, and those DERs may sit behind the customer meter or on the distribution network depending on programme design and regulatory structure.

DOE's Liftoff reports make the scope concrete: the canonical examples are rooftop solar paired with behind-the-meter batteries, EVs and chargers, electric water heaters, smart buildings and their controls, and flexible commercial and industrial loads — all customer-sited — and DOE explicitly counts traditional demand response inside its VPP definition rather than as a parallel category. NARUC's "aggregated DERs" and EPRI's "distributed energy resource aggregations" name the same concept, which is worth knowing when three documents seem to describe three different things.

Aggregation and a virtual power plant are not the same thing, and it is worth being pedantic about the difference. IRENA defines an aggregator as a grouping of agents in a power system — consumers, producers, prosumers or any mix of them — to act as a single entity when engaging in markets or selling services to the operator.

The VPP is what that aggregator operates: the construct that makes the group behave like a traditional power plant with standard attributes such as minimum and maximum capacity, ramp-up and ramp-down. Aggregation is the commercial grouping; the VPP is the controllable plant model layered on top of it.

The control layer is what turns a list of assets into a dispatchable resource. IRENA describes the VPP as controlled by a central information technology system in which weather forecasts, wholesale electricity prices and overall supply and consumption trends are processed to optimise the operation of the dispatchable DERs in the portfolio.

INL's framing is that VPPs replicate capabilities traditionally limited to centralised generation, such as energy production and dispatch, while also enabling functions centralised plants could not provide, including demand-side flexibility and distributed storage coordination. In BESS terms, a VPP is a power plant controller whose plant is scattered and partly owned by other people.

Fleet control differs from a VPP on the commercial axis rather than the electrical one. An owner dispatching its own portfolio of grid-scale batteries through a shared optimiser is running fleet control: each plant typically holds its own interconnection agreement and market registration, or sits under the owner's balance responsible party, and no third party stands between asset and market.

The software can look identical to a VPP platform — forecasting, price response, portfolio co-optimisation across sites — but the defining VPP move is absent, because there is no aggregator interposed as the market participant and none of the risk transfer described below.

FERC Order 2222 makes the distinction concrete for the US: it establishes the DER aggregator as a type of market participant, which is exactly the role a fleet owner-operator has no need of. When a grid-scale operator brands its portfolio a VPP, check the registration paperwork; often the construct underneath is fleet control, with each asset facing the market directly.

The microgrid boundary is definitional, and worth drawing precisely because the two labels get swapped. DOE defines a microgrid as a group of interconnected loads and DERs within clearly defined electrical boundaries that acts as a single controllable entity with respect to the grid, and that can connect and disconnect from the grid to operate in island mode.

Neither defining element — the electrical boundary or the islanding capability — appears in the VPP definitions above: a VPP's assets sit behind many meters across a distribution network, and the DOE and IRENA descriptions ask the portfolio to behave like a plant in a market, never to island.

The categories overlap rather than exclude each other; DOE's 2025 Liftoff update describes a utility microgrid built with VPP-enabling technology, and a microgrid's assets can enrol in a VPP while the microgrid keeps its boundary and island capability. Classify a project by its defining properties — boundary and islanding on one side, market-facing aggregation on the other — and the label sorts itself.

Why it matters in a real grid-scale project

The clearest statement of why VPPs exist at all comes from ENTSO-E, the association of European transmission system operators. Large-scale generation and storage units are currently able to participate in balancing and wholesale markets on their own, or through their balance responsible party; smaller generators and storage units require aggregation services to reach the scale required for participation. That is the whole commercial thesis in two sentences. If your BESS is a 100 MW standalone project, you are the large unit and a VPP is a distribution channel you may not need.

If you are selling 5 MW of behind-the-meter batteries, aggregation is the only door into the market. US regulation writes the same boundary into codified text: FERC defines a distributed energy resource as any resource located on the distribution system, any subsystem thereof or behind a customer meter — storage, distributed generation, demand response, energy efficiency, thermal storage, electric vehicles and their supply equipment — so a transmission-connected grid-scale battery sits outside the DER category and participates in wholesale markets directly rather than through an aggregation.

ENTSO-E also names the risk transfer, which is the part developers tend to underprice. Aggregation services give prosumers and small generators the technology and control they lack, and the aggregator acts as the responsible party in the power and flexibility markets, which moves market responsibility off the asset owner.

Aggregation through VPPs reduces the risk to each individual generator or smaller storage unit of providing flexibility, because that risk is taken by the aggregator and diversified across several flexibility providers. Read that as: a VPP operator is selling portfolio statistics as much as it is selling megawatts, and its margin lives in the diversification.

For a battery engineer the practical consequence is that identical hardware faces a different obligation set inside a VPP. DOE's public framing is that VPPs, generally considered a connected aggregation of DER technologies, offer deeper integration of renewables and demand flexibility.

Deeper integration means an aggregator's dispatch signal, rather than the site's own schedule, drives state of charge — so cycle count, depth of discharge and the resulting degradation are set by a third party's optimiser. Any capacity warranty written against an operating envelope has to survive that, and the enrolment contract needs to bound annual cycling the same way an EPC contract bounds duty.

Typical values and standards

There is no standard to cite, and that is the finding to carry away. As of mid-2026 no VPP definition could be located in an IEC or IEEE standard, nor a formal definition published by ACER, CEER, Ofgem or NERC.

What exists is a policy definition from the US Department of Energy, relayed by Idaho National Laboratory in INL/RPT-26-89692 of January 2026, co-authored with the consultancy ScottMadden; a description in IRENA's 2019 Innovation Landscape brief on aggregators; and an ENTSO-E technology factsheet that is descriptive rather than definitional — it explains what aggregation does but never writes "a VPP is defined as".

Weigh the provenance before quoting any of it in a term sheet. The DOE sentence that INL quotes is taken from DOE's 2023 report "Pathways to Commercial Liftoff: Virtual Power Plants", so the primary source is a policy document that a national laboratory is relaying, not independent national-laboratory analysis.

ENTSO-E, which has 40 member transmission system operators from 36 countries, publishes its VPP text as a technology factsheet. IRENA is an intergovernmental agency writing an innovation brief. None of the three is a regulator and none is a standards body, so the definition that binds you is always the one written into the participation rules of the market you are bidding into.

The nearest thing to a regulatory framework is US federal, and it regulates the activity without defining the term. FERC Order No. 2222 — issued September 17, 2020, published at 85 FR 67094 and effective December 21, 2020 — requires each US RTO and ISO to carry tariff provisions that let distributed energy resource aggregations, built on the DER definition quoted earlier, participate directly in its capacity, energy and ancillary service markets, and to establish DER aggregators as a type of market participant: the codified home of the aggregator role ENTSO-E describes.

The order caps the minimum size an RTO may impose on an aggregation at 100 kW — the cap binds the aggregation, individual devices carry no federal minimum, and RTOs may set lower floors.

The order text also bars an RTO from accepting bids from an aggregation that includes customers of utilities distributing 4 million MWh or less in the previous fiscal year unless the relevant retail regulatory authority opts those customers in, so retail access stays a state-level question even after the federal door opened. Compliance filings were due by September 17, 2021; what happened next is the part to check.

Implementation is where the plan met the queue. Per the DOE-sponsored compliance tracker, CAISO — the first US ISO to implement a DER aggregation model, in 2016 — brought its Order 2222 program live on November 1, 2024, and NYISO has run an earlier-approved aggregation program since April 2024 with full compliance targeted for the end of 2026.

ISO-NE's energy and ancillary service rules for aggregations take effect November 1, 2026, with the capacity piece tied to a delayed auction redesign; PJM moved full implementation from February 2, 2026 to February 2, 2028; and MISO and SPP are pointing at the end of the decade, around 2029 to 2030. A 2020 order still phasing in through 2028 and beyond is the operative fact: before a revenue model assumes aggregated participation, confirm that the specific market's door is open, at what aggregation size, and for which products.

The scale figures are policy numbers and should be quoted with their pedigree attached. DOE's January 2025 Liftoff update, citing Wood Mackenzie's North America VPP market report, puts current capacity at roughly 33 GW across North America, revising the 2023 report's 30-60 GW estimate downward.

The department's stated 2030 goal is 80-160 GW — roughly a tripling, framed as enough to serve 10 to 20 percent of peak load and to cut overall grid costs by about $10 billion per year. A target and a consultancy estimate are useful for sizing the ambition; the market-by-market participation arithmetic above is what a project can actually build a revenue case on.

In place of a standard, the parameters that actually define a VPP sit in the market registration: minimum bid size, telemetry and metering requirements, dispatch latency, settlement basis, and who is the balance responsible party. ENTSO-E's framing points straight at those, since aggregation exists to reach the scale required for market participation and the aggregator acts as the responsible party.

Ask which market products the portfolio is registered for, what happens contractually when a share of the fleet fails to respond, and whether availability is measured per asset or across the portfolio. Those answers are the specification. The label is not.

How it shows up in specs, studies and contracts

VPPs come in two shapes and the paperwork should say which. INL, relaying DOE, notes that these aggregations can serve various grid roles: some VPPs primarily shape demand by orchestrating behind-the-meter consumption and generation, while others export electricity back to the grid.

The two settle differently, meter differently and have very different interconnection consequences. A demand-shaping VPP is a load-side product; an exporting VPP puts your battery behind an export-capable connection with all the protection, anti-islanding and grid-code obligations that implies. When a term sheet says "VPP participation", resolve that binary before pricing anything.

Read an enrolment agreement the way you would read a tolling agreement, because functionally that is what it is. Pin down dispatch rights: activations per year, depth of discharge per activation, notice period, and the state-of-charge management rules between events.

Pin down the responsibility split ENTSO-E describes, where the aggregator acts as the responsible party in the market — that phrase should map onto a specific imbalance liability clause rather than a statement of intent. Pin down the measurement point and the baseline. Then confirm the aggregator's dispatch rights fit inside your capacity warranty's throughput cap, or the two documents will contradict each other in year three.

On the technical side, the VPP platform is a supervisory layer above your energy management system and power plant controller, and the integration work is real. IRENA's description — a central IT system processing weather forecasts, wholesale prices and supply and consumption trends to optimise dispatchable DERs — implies a telemetry feed out and a setpoint feed in, at a cadence the market product dictates.

DOE's 2025 Liftoff update records how wide that cadence spread runs in the US: NYISO requires six-second telemetry from every DER asset of at least 100 kW regardless of the service provided, while PJM allows one-minute scans for resources not providing regulation and exempts DERs under 10 MW from telemetry reporting entirely.

Specify the protocol, the point list, the update rate, the failure behaviour when the link drops, and the precedence between VPP setpoints and local limits such as state-of-charge floors, thermal derates and grid-code obligations. Local protection must always win that argument.

Common pitfalls

The first pitfall is assuming "VPP" has a settled meaning. It does not — as of mid-2026 the term rests on policy and industry descriptions, and two counterparties can use it for very different products. One means an aggregated demand-response programme with no export at all; another means a fleet of grid-scale batteries bidding into wholesale energy and balancing markets.

Write out what the portfolio actually does, which markets it clears in, and which entity is the responsible party. If a document uses "VPP" as though it were a defined term, it is either quoting a specific market's participation rules, in which case cite them, or it is marketing.

The second is treating aggregation and a VPP as synonyms. Aggregation, in IRENA's words, is the grouping of agents to act as a single entity in markets; the VPP is the operated construct that makes the group present minimum and maximum capacity, ramp-up and ramp-down like a plant.

Plenty of products marketed as VPPs are the former without the latter — a billing arrangement across a set of assets, with no central optimisation and no ability to hold a ramp rate. Ask for dispatch performance data, activation success rate and response time against the market product's requirement, before accepting that the portfolio behaves like a plant.

The third is missing where the risk actually sits. ENTSO-E is explicit that the aggregator takes the flexibility-provision risk and diversifies it across several providers, which is a benefit to the small asset owner and a concentration for the aggregator.

If you are the aggregator, your exposure is correlated failure: a common firmware fault or a shared weather event that moves the whole portfolio at once, precisely when the market needs it. If you are the asset owner, your exposure is that someone else's optimiser is spending your cycle life. Neither risk is visible in the megawatt number on the front page.

The fourth is numerical and US-specific: the two 100 kW rules. Order 2222's 100 kW caps the minimum size an RTO may impose on an aggregation; NYISO's 100 kW is a per-asset telemetry threshold. The rules are unrelated, and conflating them produces enrolment criteria that exist in neither rulebook.

Numbers in this space also go stale quickly — the PJM effective date and the headline US capacity estimate quoted above have each been revised once already — so where the defining documents are policy reports and compliance filings rather than standards, every figure needs a revision-date check before it lands in a model or a term sheet.