NEWSAR
Multi-perspective news intelligence
SRCThe Guardian - World News
LANGEN
LEANCenter-Left
WORDS435
ENT11
THU · 2026-09-24 · 17:26 GMTBRIEF NSR-2026-0924-113844
News/Launch of UK’s ‘largest AI supercomputer’ delayed by power s…
NSR-2026-0924-113844News Report·EN·Technology

Launch of UK’s ‘largest AI supercomputer’ delayed by power supply problems

The UK's planned "largest AI supercomputer" datacentre in Loughton, Essex, faces significant delays, potentially pushing its launch from next year into the mid-2030s. Nscale, the startup behind the project, is encountering power supply problems, with UK Power Networks reportedly unable to provide sufficient electricity until the early to mid-2030s.

Dan Milmo Global technology editorThe Guardian - World NewsFiled 2026-09-24 · 17:26 GMTLean · Center-LeftRead · 2 min
Launch of UK’s ‘largest AI supercomputer’ delayed by power supply problems
The Guardian - World NewsFIG 01
Reading time
2min
Word count
435words
Sources cited
5cited
Entities identified
11entities
Quality score
100%
§ 01

Briefing Summary

AI-generated
NEWSAR · AI

The UK's planned "largest AI supercomputer" datacentre in Loughton, Essex, faces significant delays, potentially pushing its launch from next year into the mid-2030s. Nscale, the startup behind the project, is encountering power supply problems, with UK Power Networks reportedly unable to provide sufficient electricity until the early to mid-2030s. This delay highlights broader issues in datacentre development, as the energy regulator Ofgem has noted a substantial backlog of datacentres seeking electricity connections. The Loughton site requires 90MW of capacity, a demand that necessitates upgrades to the wider national electricity transmission network. Nscale is exploring on-site power generation and ways to expedite the connection process.

Confidence 0.90Sources 5Claims 5Entities 11
§ 02

Article analysis

Model · rule-based
Framing
Technology
Economic Impact
Tone
Measured
AI-assessed
CalmNeutralAlarmist
Factuality
0.80 / 1.00
Factual
LowHigh
Sources cited
5
Well sourced
FewMany
§ 03

Key claims

5 extracted
01

The Loughton site requires up to 90MW of capacity, equivalent to the energy consumption of approximately 315,000 homes.

statistic
Confidence
0.90
02

Ofgem warns of a 'logjam' in electricity connection applications, with 315 datacentres queueing for 73GW of demand.

statisticOfgem
Confidence
0.90
03

The project, announced in 2025 as the country’s largest AI supercomputer, was initially planned for a 2027 launch.

factual
Confidence
0.90
04

UK’s largest AI supercomputer project in Loughton, Essex, faces delays due to power supply problems.

factual
Confidence
0.90
05

UK Power Networks has informed Nscale that the local grid cannot supply sufficient power until the early to mid-2030s.

factualUK Power Networks (reported)
Confidence
0.80
§ 04

Full report

2 min read · 435 words
A huge datacentre project hailed by the UK government will miss its launch date next year and could be delayed into the mid-2030s.The site in Loughton, Essex, was described as the country’s largest AI supercomputer when it was announced in 2025, but power supply problems mean it now faces a lengthy wait before coming online.Doubts had already emerged about the scheme after The Guardian revealed in March that the planned location was still an operational scaffolding yard.Now there are further concerns about the project due to its inability to obtain enough power before its planned 2027 debut. The Loughton supercomputer is being built by Nscale, a UK-based startup seeking a $35bn (£26bn) flotation this year.UK Power Networks (UKPN), the company that will connect the site to the local power grid, has reportedly told Nscale the grid will not be able to supply sufficient power to the datacentre until the early to mid-2030s.The power delay was first reported by Morning Intelligence, an AI industry newsletter.Nscale said it was “committed to the project” but the failure tofind a quick fix to the power supply issue underlines how much energy has become a hindrance to datacentre development. The scaffolding yard site has now been cleared by Nscale, however. The company is also investigating whether it can generate power onsite and whether it can speed up the connection process.Keir Starmer’s government cited the Loughton project when announcing the UK’s AI strategy in 2025, describing the site as the “largest UK sovereign AI datacentre”. Datacentres are the central nervous system of AI technology such as chatbots.The scaffolding company, which was operating on the site. Photograph: Martin Godwin/The GuardianOfgem, the energy industry regulator, has warned of a logjam in applications for electricity connections. In June, it revealed there were 315 datacentres queueing to connect to the National Grid, representing 73GW of demand – compared with the peak energy demand for the entire country of 45GW.The Loughton site is seeking up to 90MW of capacity, equivalent to the energy consumption of about 315,000 homes.UKPN said it could not comment on an individual project. Schemes with large energy demand also require work to be done at a National Grid level, in order to ensure that energy demands can be met properly.“In general, many datacentres, whilst they will connect to our distribution network, are awaiting works on the wider national electricity transmission network before they can begin operation,” said a UKPN spokesperson.National Grid said: “We’re continuing to work closely with NESO [the government-owned National Energy System Operator], distribution network operators and customers to identify opportunities to accelerate connections where possible and support economic growth.”
§ 05

Entities

11 identified
§ 06

Keywords & salience

10 terms
datacentre
1.00
power supply problems
1.00
ai supercomputer
1.00
uk government
0.80
nscale
0.70
energy demand
0.70
uk power networks
0.60
ai strategy
0.50
electricity connections
0.50
loughton
0.40
§ 07

Topic connections

Interactive graph
Network visualization showing 51 related topics
View Full Graph
Person Organization Location Event|Click node to navigate|Edge numbers = shared articles