The scale of the problem is structural. By late 2024, the combined UK connection queue across transmission and distribution networks had passed 700 gigawatts, much of it speculative, unfunded projects clogging positions ahead of genuinely ready developments.

NESO’s TMO4+ reform, backed by Ofgem, is working through that backlog by prioritising projects with land rights, consents and financial commitment already in place. Data centres were classified as critical national infrastructure in September 2024. Ofgem followed with a further consultation in July 2026, targeting queue-management milestones specific to the sector.

A new grid connection agreement still sits in a queue for years. Cogeneration for data centre applications matters here because it changes which queue, if any, a project has to join. This article sets out that mechanism, and what it means for specifying a plant.

What cogeneration delivers: CHP efficiency explained

Cogeneration efficiency is calculated by measuring the combined electrical and thermal output against total fuel input, and the two need separate scrutiny, since each is measured and used differently on site. A gas engine generating electricity alone typically converts around 40 to 45% of fuel energy into power and the rest is lost as heat.

Combined heat and power recovers a meaningful share of that waste heat from the jacket water and exhaust streams, raising overall fuel utilisation substantially above electrical-only generation. UK qualifying CHP efficiency is assessed under the CHP Quality Assurance (CHPQA) programme, against Power Efficiency and Quality Index thresholds: 20% for the former, and for the latter, 95 during Initial Operation, rising to 100 once a scheme reaches Annual Operation. Any figure quoted externally should be checked against the scheme’s own certification.

The distinction matters for data centre applications specifically. Electrical output runs the IT load; recovered heat can offset cooling energy where it drives absorption chillers, or serve auxiliary heating loads elsewhere on site. Therefore, CHP efficiency figures must state what counts as useful heat and how it was measured, without that detail, a quoted figure can overstate real site performance.

Combined heat and power generation: the system architecture

A cogeneration plant for a data centre is built around a reciprocating gas engine driving an alternator, with heat recovery equipment attached to the engine’s jacket water circuit and exhaust gas stream. Simple in principle. The alternator supplies electrical power either directly to the data hall or in parallel with the grid connection, depending on the site’s configuration. Recovered heat, meanwhile, is typically routed through a plate heat exchanger into a hot water circuit.

That hot water can serve two purposes on a data centre site. It can meet conventional heating demand for office or auxiliary space, or drive an absorption chiller to generate chilled water for precision cooling. The second route is what turns cogeneration into combined cooling, heat and power- the configuration most relevant to high-density server halls with a year-round cooling load.

System sizing depends on matching engine output and heat recovery capacity to the facility’s actual electrical and thermal demand profile, not to headline IT load alone. Oversizing wastes fuel utilisation gains; undersizing forces reliance on grid or diesel backup during peak periods. This is engineering work that has to happen before procurement, not after.

Gas CHP for data centres: advantages and disadvantages

The advantages are concrete: grid exposure, fuel efficiency, and phased capacity. Onsite generation reduces exposure to grid connection timelines, converts a larger share of fuel energy into useful output than grid-supplied power plus separate cooling, and can be deployed in modular stages that match a data centre’s phased build-out. Reciprocating gas engines also start and reach full load faster than most alternative onsite generation technologies which is especially useful when facilities need rapid capacity additions.

The disadvantages need equally specific treatment. A CHP plant depends on continuous, secure gas supply, and any interruption to that supply is a direct risk to onsite generation capacity. Scheduled maintenance intervals on reciprocating engines are more frequent than for some alternative technologies, requiring redundancy in the design. Physical footprint for engines, heat recovery equipment and any absorption chillers is larger than an equivalent electrical-only backup system- which constrains options on dense or brownfield sites.

Emissions accounting is a further consideration site teams should not gloss over. Natural gas combustion still produces carbon dioxide, and any sustainability case for CHP needs to weigh fuel utilisation gains against total emissions compared with grid electricity from a decarbonising supply mix. Dual-fuel or hydrogen-ready engine specifications can mitigate this over the asset’s lifetime. That mitigation depends on a deliberate design decision made at the specification stage.

Energy independence in practice: embedded generation and grid Codes

Genuine energy independence depends on how a cogeneration plant is classified and connected. A plant connected as embedded generation, behind the site’s own meter, sits on the distribution network in a way that avoids the same transmission-level generation connection queue faced by a new grid-scale asset. That classification is what allows deployment to proceed on an engineering and construction timeline, rather than a queue position.

Connecting any generating plant to the UK network, whether operating in parallel with the grid or capable of islanded operation, requires compliance with the relevant Engineering Recommendation- currently G99 (Issue 2, March 2025), covering generation connected to distribution networks in Great Britain. Compliance covers protection settings, synchronising equipment, and the conditions under which the plant can disconnect from or reconnect to the grid safely.

Two operating modes matter here. Grid-parallel operation lets the CHP plant run continuously alongside the grid connection, reducing imported load and allowing a controlled export or zero-export arrangement. Islanded operation allows the plant to disconnect and continue supplying the site independently during a grid outage- the scenario most relevant to uptime-critical facilities. Both require engineering sign-off against G99 before commissioning, and the distinction should be specified early.

Why this removes the multi-year wait as a blocker to going live

A new grid connection agreement for additional demand or generation still joins the queue described above, currently running to a decade or more in many regions. An embedded generation asset behind the meter, sized to the site’s own load, avoids that same connection agreement and the same competition for transmission capacity.

This does not remove regulatory steps entirely. A site still needs planning consent, environmental permitting for combustion plant, gas supply agreements, and G99 compliance sign-off before it can operate- but those processes run on an engineering and construction timeline measured in months, set against a connection queue measured in years. For a data centre operator weighing options against a seven-to-thirteen-year London wait, that difference decides whether a facility goes live this decade or the next.

Modular, factory-built CHP plants shorten that timeline further. Containerised systems, assembled and tested before delivery, reduce onsite construction time relative to a custom-built plant, and capacity can be added in phases as a data hall scales. A pre-tested, factory-assembled unit removes a coordination step that a site-built system, assembled from separately sourced components, still has to work through.

Specifying gas CHP plants for a data centre: engineering considerations

Sizing has to account for phased IT load growth rather than day-one demand alone, because a CHP plant undersized against future load forces early reliance on grid or diesel top-up. Undersizing is the costly mistake. Redundancy design should follow the facility’s own reliability target, N+1 or N+2 across engines, rather than assuming a single plant meets uptime requirements. Fuel supply security, including contracted gas volumes and any backup fuel arrangement, needs the same scrutiny as the mechanical specification.

How a CHP plant responds during transition events must be modelled before procurement. This means factoring in integration with existing UPS and battery energy storage assets on site from the outset. Grid code compliance sign-off under G99 matters too- it should be scoped as an early project milestone, since it gates both islanded operation and any parallel export arrangement. Each of these decisions changes the plant specification, so they belong in the engineering brief before a supplier is selected.

The post Cogeneration For Data Centres: How Onsite CHP Bypasses The Grid Connection Queue first appeared on BioEnergy Consult.