Nvidia Is Looking at Old Telephone Buildings for AI Data Centres as Demand for Computing Power Grows

The exploratory plan could provide access to about 1 gigawatt of power across up to 20 states, but costs, permits and customer commitments remain unclear

Nvidia
Nvidia is exploring plans to convert more than 130 Brightspeed telecom buildings into AI data centres, potentially unlocking 1GW of power capacity Mariia Berezovsky/Unsplash

Nvidia has proposed turning more than 130 telecommunications buildings operated by Brightspeed into AI data centres across the United States, The Information reported on 5 November 2026.

The plan, discussed with private equity firm Apollo Global Management, could give the chipmaker access to an estimated 1 gigawatt of power capacity.

The proposal comes amid growing opposition to large data centre developments. Some winning candidates in recent US gubernatorial elections, including in Texas, have called for moratoriums on new facilities.

Nvidia Sees An Opportunity In Brightspeed's Telecom Network

Brightspeed operates more than 130 buildings originally used to manage voice calls and data traffic. The broadband provider was formed in 2022 after Apollo acquired parts of Lumen Technologies' local telephone and broadband businesses for $7.5 billion (£5.59 billion).

Nvidia first raised the idea with Apollo executives, who asked the chipmaker to submit its plans in writing to Brightspeed. Nvidia then sent a proposal requesting technical information and visits to several sites, The Information reported, citing people familiar with the discussions.

The proposed network could span as many as 20 states, although it remains unclear how many buildings Nvidia wants to convert. Locations mentioned in the report include Dothan, Alabama, Killeen, Texas, and Mansfield, Ohio.

Nvidia estimates the buildings could offer around 1 gigawatt of power, although their suitability for AI data centres remains unconfirmed.

The company has discussed starting retrofits in early 2027, potentially installing Vera Rubin servers and Groq technology, but no construction schedule has been confirmed.

The Proposed AI Data Centres Would Focus On Inference

Nvidia has described the proposed facilities as an 'inference network', designed to run already-trained AI models rather than train new ones. Smaller sites could help the company and potential customers tap into pockets of available electricity without relying on massive data centre campuses.

There are practical obstacles. Older telecommunications buildings may need specialised cooling systems and electrical upgrades to support modern AI servers.

Those works could still require local permits, although the original report said they could involve fewer hurdles than building new facilities.

Zac Smith, chief executive of AI networking company Datum, told The Information that older telecom properties can unlock underused electricity. He also warned that they can sometimes be more expensive to operate because of ageing electrical equipment.

Apollo's Role And The Unanswered Customer Question

Apollo's involvement extends beyond its connection to Brightspeed. The investment firm recently agreed to buy tens of billions of dollars' worth of Vera Rubin chips, which it plans to lease to SpaceX.

That arrangement does not establish that SpaceX will use the proposed Brightspeed facilities. Potential customers could include SpaceX, Anthropic and OpenAI, The Information reported, but it was unclear whether Nvidia or Apollo had approached them about this project.

Brightspeed's financial position is another sensitive part of the story. People familiar with the effort told The Information that the company was reportedly nearing bankruptcy. This is not confirmation of a bankruptcy filing.

The proposal remains exploratory. Nvidia has requested information and site visits, but no final agreement, confirmed customer list or approved conversion programme has been disclosed.

For now, questions remain over how many buildings are suitable, how much upgrades will cost and how much power they can provide before the project moves forward.