★ INSERT COINNOW PLAYING: VENTURESHIGH SCORE: $100M ARR★ NEW STAGE UNLOCKED: ABOUT MEPRESS START★ DEMO DAY 04:00:00
★ INSERT COINNOW PLAYING: VENTURESHIGH SCORE: $100M ARR★ NEW STAGE UNLOCKED: ABOUT MEPRESS START★ DEMO DAY 04:00:00
◀ BACK TO FEED
NEWSDEEP TECHJUL 13, 2026

Nscale’s £2B data centre runs into the AI power bottleneck

Nscale’s Essex data-centre project shows that grid access, not just chips or capital, is becoming a defining constraint on AI infrastructure growth.

Nscale’s £2B data centre runs into the AI power bottleneck

AI data centres can secure financing and chips long before they secure enough electricity to operate.

What happened

Nscale’s planned £2B data centre in Essex has reportedly been told that its proposed 90MW grid connection will not be ready for the site’s intended 2027 opening.

The company is discussing alternative power options, including solid-oxide fuel cells supplied by Bloom Energy, to reduce its dependence on the delayed grid connection.

The site is expected to provide Nvidia GPU capacity to major customers, including Microsoft, making the power delay commercially significant rather than a routine construction issue.

Why it matters

Data-centre development is increasingly constrained by power availability, connection queues and local grid capacity.

A project can have land, capital, customers and computing hardware lined up while still being unable to operate on schedule because the electricity system cannot support it.

Alternative on-site generation may help bridge that gap, but it also adds cost and complexity.

The bigger picture

The AI infrastructure race is becoming an energy-planning race.

Developers may need to pair data centres with dedicated generation, storage or long-term power agreements instead of relying entirely on existing grids. That shifts AI infrastructure closer to the economics of heavy industry, where energy access can determine which projects scale and which remain delayed.

#NSCALE#AI DATA CENTRES#GRID INFRASTRUCTURE#NVIDIA