Key points
- Microsoft is no longer the anchor customer for the planned Loughton, Essex AI data center.
- Grid operators told builder Nscale that sufficient power may not be available until the year 2030.
- The facility was originally scheduled to begin operations in early 2027 with 50 megawatts of capacity.
Microsoft has withdrawn its commitment to serve as the anchor customer for a major artificial intelligence data center planned for Loughton, Essex, according to reports from The Telegraph and Network World. The decision stems from significant delays in securing an electricity connection for the site. Grid operators have informed Nscale, the company building and operating the facility, that the required power supply will likely not be available until 2030. This timeline renders the original deployment schedule unworkable for Microsoft.
By the numbers
- 23,040: The number of Nvidia GB300 processors the facility was intended to house.
- 50 MW: The initial planned computing capacity of the Loughton site.
- 90 MW: The potential scaled capacity of the data center using the processor array.
What the figures mean
The scale of the hardware deployment indicates a high-density AI computing environment. With nearly twenty-four thousand processors planned, the facility was designed for substantial workloads. The shift from a 2027 start date to a 2030 power availability window creates a multi-year gap. This delay impacts the immediate utility of the infrastructure for tenants requiring current-generation AI capacity. Microsoft was the only company that had publicly committed to using the site.
What should you do?
Managed service providers should review their own infrastructure roadmaps for similar utility dependency risks. Teams need to verify the power capacity guarantees in contracts with data center builders. It is prudent to ask suppliers about their contingency plans for grid delays. Organizations should assess whether their AI workloads can tolerate multi-year deployment shifts. Stakeholders must understand that a single site issue may not reflect broader regional stability.
Operational impact
When an anchor tenant exits a project due to infrastructure constraints, service providers must evaluate their exposure to similar delays. Teams usually need to confirm whether their service level agreements include clauses for utility failures. Check whether your suppliers have diversified power sources or backup generation plans. Ensure that internal risk registers reflect the possibility of extended construction timelines for new facilities. Communicate these potential delays to clients who may depend on the promised capacity. Do not assume that a delay at one site affects all projects in the same region.
What to do and how to stay safe: Microsoft
- Review current contracts with data center providers for clauses regarding power availability and construction delays.
- Assess whether your AI or high-performance computing workloads can be shifted to existing facilities with confirmed power capacity.
- Monitor official statements from infrastructure builders regarding alternative power sources for delayed sites.
- Consult with your procurement team to identify backup suppliers that can offer immediate capacity if primary plans stall.
Step-by-step guide: Shadow AI Explained: 8 Questions Answered for Secure Operations
General security guidance from the BitShield News newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
Why did Microsoft withdraw from the Essex data center project?
Microsoft withdrew because grid connection delays pushed the expected power availability into the 2030s, making the original 2027 schedule unworkable.
Who is building the Loughton data center?
Nscale, a British AI infrastructure company, is the owner and builder of the data center facility.
Will this affect Nscale's relationship with Microsoft?
No, a Microsoft spokesperson stated the decision relates to a single site and that their partnership with Nscale remains strong and long-term.
