Novexa News

London Data Centre Plan Faces Million-Tonne Carbon Test

A proposed hyperscale data centre in North Ockendon could produce more than one million tonnes of carbon emissions annually, planning documents estimate.

Abdul BasitPublished August 21st, 2026 6:35 PMUpdated August 24th, 2026 7:25 PM4 min read
Large data centre campus beside green fields and trees on the edge of London

Image credit: AI-generated editorial illustration by Novexa News

A proposed hyperscale data centre on London's eastern edge is facing scrutiny after planning documents estimated that the development could be responsible for more than one million tonnes of carbon dioxide emissions each year.

The East Havering Data Centre Campus is planned for North Ockendon, within the London borough of Havering. The project would occupy about 218 hectares of green-belt land and could become one of Europe's largest data-centre sites if councillors grant approval.

The application arrives as demand for cloud computing and artificial intelligence infrastructure is rising rapidly. It also creates a difficult planning question: how should cities weigh digital capacity and investment against electricity demand, land use and legally binding climate goals?

What the emissions estimate represents

Reporting on the planning documents compared the projected annual footprint with about 27,000 flights between London and New York. That comparison is intended to make a very large carbon number easier to understand, but the underlying estimate deserves careful treatment.

The figure is a forecast associated with a proposed development, rather than a measurement from an operating facility. Actual emissions would depend on how much computing capacity is built and used, the efficiency of servers and cooling systems, the electricity mix, backup generation and any credible low-carbon supply contracts.

Those variables do not make the estimate irrelevant. They show why planners need transparent assumptions and enforceable conditions. A project can promise future efficiency while still creating near-term demand that must be met by the power system.

Why AI infrastructure uses so much power

Data centres operate around the clock. They consume electricity to run processors, store information, move data across networks and keep equipment within safe temperatures. Large AI systems add intensive computing workloads, especially during model training and high-volume use.

The environmental footprint extends beyond electricity. Facilities require water or alternative cooling systems, backup power, substations, transmission connections and replacement hardware. Construction also carries embodied emissions from concrete, steel and other materials.

Supporters argue that modern hyperscale facilities can be more efficient than smaller, older server rooms and that digital services support economic activity across many sectors. The relevant comparison is therefore not simply data centre versus no data centre. It is whether the proposed design delivers necessary capacity with the lowest realistic environmental cost.

The green-belt question

The use of 218 hectares of green-belt land is likely to be central to Havering council's decision. Green-belt policy is intended to limit urban sprawl and preserve open land around cities. Development can be approved in exceptional circumstances, but applicants must make a strong case for the location and public benefit.

Planning authorities will need to consider landscape impact, biodiversity, transport, construction disruption and power infrastructure alongside emissions. They will also need to examine whether less damaging sites or designs could meet the same need.

The debate should not be reduced to a choice between technology and nature. Location, grid access and design determine whether a data centre supports climate-compatible digital growth or locks in avoidable pressure on land and energy systems.

What a credible approval would require

If the campus proceeds, meaningful safeguards would need more than broad sustainability language. Planning conditions could require detailed energy reporting, strict efficiency targets, limits on diesel backup use, heat-reuse planning, habitat protection and milestones tied to clean electricity.

Independent monitoring would matter because computing demand can change quickly after permission is granted. A campus built in phases may ultimately host different equipment and workloads from those assumed at the application stage.

The proposal also highlights a national policy gap. Britain wants investment in AI and cloud infrastructure, but local councils are often left to assess projects whose energy implications extend beyond their boundaries. Clearer national standards could help prevent inconsistent decisions and give developers certainty about what is expected.

Havering council's eventual decision will concern one site, yet its significance is wider. As AI becomes embedded in everyday services, the physical infrastructure behind it is becoming impossible to treat as weightless. The North Ockendon plan puts that reality, and its carbon cost, directly before the planning system.

Source links

Comments

No approved comments yet.

Related Articles