Markets and strategy stage summary

Flexibility used to be a set of bilateral arrangements. It is becoming a market with named institutions. Elexon was appointed market facilitator in July 2024, Ofgem's policy framework decision followed in June 2025, and the role has been operating from early 2026. Beside it sits FMAR, Flexibility Market Asset Registration: an Ofgem programme with Elexon appointed delivery body on 7 March 2025, building a one-stop flexibility asset register due to launch in 2027. Getting the expansion right matters, because FMAR is about registering assets, not reviewing arrangements, and the two readings point at completely different work.

Registration is the precondition for stacked revenues. An asset that wants to earn from more than one market has to be identifiable, with its capabilities and its commitments visible to everyone who might dispatch it, otherwise the same megawatt gets sold twice. That is also why primacy exists: the ENA Open Networks primacy rules set out how conflicts between NESO and DNO dispatch requests are resolved, and in data terms a primacy conflict is a pair of overlapping instructions against one registered asset. The Embedded Capacity Register is the public record of distribution-connected resources, and storage co-location is the awkward edge case, because the metering arrangement decides how one site's flows are split and settled.

The reserve products moved underneath all of this, and every move rewrote a schema. Balancing Reserve went live in March 2024. Quick Reserve arrived from December 2024, with a second phase in September 2025, replacing Fast Reserve. Slow Reserve followed from March 2025, and STOR ended at 05:00 on 1 April 2026 after two decades. The Demand Flexibility Service became a year-round NESO service on 27 November 2024 and went bi-directional and zonal with a 0.1 MW threshold on 9 April 2026. A provider who missed any one of those changes was submitting data in a format nobody reads.

Behind the meter, the interesting data is now produced by appliances rather than by the meter. PAS 1878:2021 sets the architecture: energy smart appliances, Customer Energy Managers that coordinate them, and demand side response service providers that sell the resulting flexibility. PAS 1879:2021 adds the code of practice for demand side response operation. The argument the module makes is blunt: a standard is what decides whether a heat pump's operating data is portable between service providers or locked inside one manufacturer's cloud, and that decision is taken years before any consumer notices it.

When control signals start moving megawatts, the people sending them get regulated. The DESNZ Smart Secure Electricity Systems programme is the enduring governance response, and Ofgem's load control licensing consultation ran from December 2025 to February 2026. The direction is to bring load controllers inside the licensing perimeter rather than leave aggregated domestic control as an unregulated activity, which turns a software function into a regulated one with data obligations attached.

Tariff data is becoming public infrastructure on a dated timetable. Standardised machine-readable tariff data arrives through supply licence condition 11C and REC Schedule 35 from 18 May 2026, public tariff pricing APIs follow from 18 February 2027, and consumer-specific tariff sharing lands in November 2027. Ordering those three matters, because public pricing and consumer-specific sharing are different products with different consent positions. The stage closes the loop with a worked half-hourly costing: the same appliance run priced against half-hourly unit rates at different times of day is what connects device flexibility to tariff data in a form a consumer can actually see.

Gas has its own spine, and it was built early. Xoserve is the central data services provider under the Uniform Network Code. UK Link, rebuilt through Project Nexus in 2017, is the retail backbone. Gemini handles National Transmission System capacity, nominations and balancing, operated by Xoserve for National Gas. The Gas and Electricity Enquiry Services handle enquiries, and they replaced the older Data Enquiry Service on 18 July 2022, so anyone still pointing at the DES is pointing at a retired service. The uncomfortable observation is that gas centralised its data services roughly a decade before electricity attempted the same thing.

Allocation on the gas side works without half-hours. Daily metered sites are read daily; non-daily metered sites are estimated and allocated, which is the gas equivalent of the profiling that MHHS is removing from electricity. Underneath sits the two-input measurement the course has carried since stage 1: metered volume converted with calorific value, through Volume multiplied by CV multiplied by 1.02264 and divided by 3.6, to give the kilowatt hours a bill is written in. Demand estimation fills the gap where meters do not report daily.

Unidentified Gas is what happens when identifier quality and read quality problems meet a mass balance equation. The gas that entered a distribution zone has to equal the gas allocated to meter points, and the residual is UIG. It rose sharply after Nexus, and the industry's answer is the Allocation of Unidentified Gas Expert, with Engage Consulting reappointed in August 2025 to produce the annual statement and the weighting factors that spread the residual across user classes. The worked example is deliberately small: one mis-read Class 4 meter moves costs for everyone else in the zone, which is the clearest demonstration in the whole course that data quality is not an internal housekeeping matter but a transfer of money between parties.

The honest comparison is dimension by dimension, not a single verdict. On settlement granularity GB is the coarsest of its peers: Australia's National Electricity Market moved to five-minute settlement in October 2021, US ISO and RTO markets settle at five minutes, the EU adopted fifteen-minute settlement in June 2025, and GB stays at thirty minutes even after MHHS completes. On hub design, Norway's Elhub is the single-hub benchmark, where one system carries meter data, switching and consumer access behind a single API, while GB integrates across several platforms. Denmark's Energinet DataHub is the second benchmark of the same family, holding around 3.3 million metering points in a supplier-centric model.

On consumer access, Australia's Consumer Data Right has been operational for energy since 2020, with an accreditation regime and standardised APIs, while GB's Smart Data Scheme under the Data (Use and Access) Act 2025 is not expected to be operational until 2027 to 2028. North America takes the voluntary route with Green Button and the NAESB ESPI standard. The EU is codifying the same access problems in law: Implementing Regulation (EU) 2023/1162 has applied since 5 January 2025, and the Common European Energy Data Space is advancing on Blueprint v3.0 from September 2025.

GB leads on other dimensions, and the module says so. The presumed-open principle in Ofgem's data guidance is more legally enforceable than most international equivalents, the DCC-operated national smart metering WAN is unusual in scale, and CIM standardisation across the fourteen DNO licence areas since November 2025 is among the more advanced distribution-level deployments anywhere. GB has also made a deliberate architectural choice: federation rather than consolidation, with the Data Sharing Infrastructure acting as a discovery and trust layer over platforms that stay where they are, instead of an Elhub-style rebuild around one system. The right exam answer names the dimension before it names the winner.

The transformation is not one programme, it is seven running at once, each with a different owner and a different maturity. The DIP is live as the market messaging layer. MHHS is mid-migration, with meter points moving to May 2027. The Consumer Consent Solution is in delivery, with trust framework partners appointed in May 2026. The Data Sharing Infrastructure is coordinated by NESO to 2028 with no fixed launch date. FMAR delivers its register in 2027. The Smart Data Scheme awaits its implementing regulations. Clean Power 2030 is the policy destination the other six serve. Matching each programme to the right owner is the exam-grade skill, and FMAR is where most people get it wrong.

Scale is the reason the infrastructure had to change rather than merely grow. Elexon expects MHHS to generate roughly 500 billion meter readings a year, which turns settlement from a batch process into a continuous data management problem, with validation and aggregation running close to real time. Quote the source-backed figure rather than deriving one from the meter estate, because the settlement denominator is more specific than the installed base.

AI and digital twins sit on top of that foundation and inherit its weaknesses. Demand forecasting, carbon intensity prediction across fourteen regions with forecasts more than 96 hours ahead, and predictive maintenance are already operational. Digital twins of the network are emerging incrementally on CIM-standardised models with live operational data overlaid. None of it improves faster than the data underneath it, which is the stage's most portable judgement: an AI model trained on fragmented and inconsistent data returns fragmented and inconsistent predictions, so the seven programmes are the precondition rather than the alternative. Assessed against the 2030 vision, some outcomes are on track and others are at risk, and the risk is nearly always a dependency between programmes rather than a failure inside one.

Governance with teeth needs three things: clear accountability, defined standards, and enforcement that goes beyond voluntary compliance. In GB those map onto the Energy Digitalisation Framework's Board, the Data Best Practice standard and the Data Coordination Function. The architecture runs from primary legislation at the bottom, through licence conditions and the framework bodies, up to the instruments that hold organisations to account. Below the Board sits the Digitalisation Delivery Group for operational coordination, and below that four domain coordinators: Core Energy Data with NESO and Consumer Data with RECCo are confirmed, while Elexon's designations for Behind-the-Meter and Metering data remain provisional.

The domains overlap on purpose, which is why the Cross-Domain Technical Working Group exists. A single smart meter reading is behind-the-meter data, metering data, and potentially consumer data all at once, depending on what is done with it. Coordination is therefore less about drawing clean boundaries and more about keeping standards and APIs consistent across territory that genuinely is shared. The open design question about the Data Coordination Function is whether it holds real operational authority or remains mainly a convening forum, and that question is worth asking of any coordination body you meet.

Two instruments carry the accountability, and they do different jobs. The Digitalisation Strategy and Action Plan is forward-looking: it records what a company intends to do. The Data Best Practice Guidance and its Supporting Information, current in its 2026 edition, is a licence-backed minimum standard describing what a compliant dataset must look like today, presumed open, discoverable, documented and openly licensed. A company can hold an ambitious DSAP while failing basic Data Best Practice requirements, which is exactly why both exist. Follow one commitment end to end and the chain is legible: the DSAP carries the promise, Data Best Practice defines the standard, the published dataset is the evidence, and RIIO price control determinations supply the funding and the re-opener that make the promise cost something to break.

The Common Information Model is the shared vocabulary for describing electricity networks, and reading it as a file format is the standard beginner error. CIM defines what things mean and how they relate; RDF/XML under IEC 61970-552, or any other serialisation, is only the container. Two organisations can exchange the same model in different formats and still interoperate, because the meaning survives the packaging. The stack has four layers: the vocabulary of classes, the profiles that constrain the vocabulary for a purpose, the exchange documents that carry actual model data, and the validation, with SHACL shapes, that decides whether a document is acceptable. The mRID, the master resource identifier, is what lets two parties agree they are describing the same physical asset.

The standard splits by domain. IEC 61970 covers transmission, IEC 61968 covers distribution with Part 11 supplying the distribution classes DNO models need, and IEC 62325 covers markets. CGMES 2.4.15 and 3.0 coexist in practice, and the GB LTDS profiles are deltas from CGMES 3.0 under BSI-led governance rather than a separate model. The nine LTDS artefact groups, EQ, SC, GL, SSH, TP, SV, SYSCAP, DL and SCR, are modular and combine into cases: an Existing Fault Level case draws EQ, SC and SCR together for protection coordination and for judging whether a new connection would push fault levels past their limits.

Delivery is live and dated, which is what makes this module practical rather than theoretical. Ofgem's derogation letter of 13 May 2026 keeps Stage 2 publication at 29 May 2026, moves SCR out of SYSCAP, defers most future-year models to Stage 3 and moves Stage 3 production to 15 October 2026. GC0139 is developing in parallel, and that parallelism is the named risk: if the Grid Code data requirements and the CIM profiles diverge, DNOs face conflicting obligations and duplicated effort. The same CIM-shaped thinking, agree the meaning before you agree the transport, is what the Data Sharing Infrastructure needs if federated discovery is going to work across the sector.

The planning ecosystem is a hierarchy of questions. The Future Energy Scenarios, published by NESO, answer how much of each technology GB needs across four pathways to 2050. The Strategic Spatial Energy Plan answers where it should go, with pathway options going to the Secretary of State in summer 2026 and the final plan due in autumn 2027. Regional Energy Strategic Plans turn the spatial answer into decisions a region can act on: the first transitional RESP was published on 30 January 2026, the methodology follows in summer 2026, and eleven full RESPs are due by the end of 2028. The Centralised Strategic Network Plan carries the transmission build, and ETYS bridges the ten-year view in between.

The distribution planning products sit beside the national ones and are easy to misattribute. Distribution Future Energy Scenarios are produced by each DNO, usually with Regen, and published as open data. They are a DNO product, not an Ofgem one, and confusing the two is the fastest way to lose credibility in a planning conversation. Alongside DFES sit Network Development Plans and the quarterly capacity heatmaps, which together are what anyone building a local flexibility or connections case actually reads.

Connections reform is where the planning data gets tested. Ofgem approved the TMO4+ package, modifications CMP434 and CMP435, on 15 April 2025, replacing first-come-first-served ordering with evidence-gated entry assessed on project readiness, alignment with the scenarios, contribution to Clean Power 2030 and deliverability, with the aim of reducing a queue of about 770 GW to roughly 381.5 GW. Gate 2 processing ran through 2025, with offers for 2030-aligned projects issued into early 2026. The dependency is the point of the whole stage: evidence-gated connections need scenarios at sub-regional granularity, CIM network models with capacity data, FMAR for flexibility visibility and the spatial plan for priorities. Break any one of those data links and the policy above it degrades quietly rather than failing loudly.