Nvidia is reportedly in talks to convert telecommunications infrastructure into artificial intelligence data centers. The company pitched the concept to Apollo Global Management, the parent firm of Brightspeed, according to Tom's Hardware citing The Information.reportedNvidia wants to turn old telephone company buildings into data centers for AI.Nvidia proposes leveraging Brightspeed's 130 facilities for 1GW capacity to mitigate greenfield data center delays.
Evidence: reported
Nvidia wants to repurpose over 130 old telco buildings as tiny data centers, report claimsaifleet-tomshardware
What was released or claimed
Apollo Global Management asked Nvidia to submit written plans to the internet provider. Nvidia complied with this request and has now asked for additional information and site visits. The proposal involves repurposing existing buildings rather than constructing new facilities from scratch.
Who it affects
Brightspeed, the internet and telecommunications provider, is the primary entity involved. Apollo Global Management, its parent company, is also central to the decision-making process. The plan could impact local governments that issue permits for these sites.
What is still unknown
It remains unclear who will own the AI chips installed in these facilities. The final agreement on terms and ownership between Nvidia, Apollo, and Brightspeed has not been disclosed. Specific timelines for implementation are also not provided in the reports.reported
What to make of it
This strategy addresses delays and community opposition associated with building large data centers. Repurposing existing telecom buildings may face less regulatory red tape. It offers a faster path to deploying compute capacity compared to traditional construction methods.
What buyers should weigh
Buyers should consider the speed of deployment offered by repurposed infrastructure. Market position may shift as vendors seek alternative site options. Vendor choice could expand to include telecom assets for AI workloads. Cost structures may differ from traditional data center builds.
Why it matters
- Repurposing existing infrastructure could accelerate AI deployment timelines significantly.
- Telecom assets offer an alternative to facing community backlash against new builds.
- Apollo's existing GPU orders highlight the urgency of finding rack space.
- The model may influence how other vendors approach data center expansion.
Topics: Nvidia, Brightspeed, Apollo Global Management, data centers, AI infrastructure
Updated
The strongest doubt
Counterpoint from the sources
The sites require new cooling systems and permits, meaning they are not plug-and-play solutions despite existing power infrastructure.
Ask the article
Local keyword search of this article; no generated answers.Did this article give you the evidence you needed?
Your choice stays on this device. No public tally.Questions readers ask
How many buildings are involved in this plan?
The plan involves approximately 130 Brightspeed buildings located across the United States.
What is the estimated compute capacity?
Nvidia estimates the sites could collectively deliver around 1 gigawatt of compute capacity.
Can the sites run the servers immediately?
No, Nvidia would still need to install cooling systems and other supporting architecture.
Connected stories
Reading receipt
- Nvidia pitched converting Brightspeed facilities to Apollo Global Management executives.
- The plan could yield 1 gigawatt of capacity across 130 U.S. sites.
- These sites could house over 2,200 Supermicro NVL72 racks.
Still open: Who will ultimately own the installed Nvidia chips?
Still open: Will this model be adopted by other telecom providers?