Key Takeaways
- SoftBank investing $3 billion to convert Ohio EV plant into OpenAI data center equipment factory
- Manufacturing of modular data centers to begin early next year
- Part of broader $18 billion joint venture announced at White House
- OpenAI aims for massive 30-gigawatt computing capacity expansion
SoftBank Group is making a massive $3 billion investment to transform an electric vehicle facility in Lordstown, Ohio into a manufacturing hub for OpenAI’s data center equipment. The Japanese conglomerate is going all-in on the AI startup at the center of the generative AI revolution.
Strategic Investment in AI Infrastructure
The repurposed plant will produce equipment for data centers in Milam County, Texas and other locations, according to The Information’s Thursday report. This move represents SoftBank’s deep commitment to OpenAI, having previously sold its $5.8 billion stake in Nvidia to fund CEO Masayoshi Son’s ambitious AI strategy.
SoftBank acquired the Lordstown facility in August for $375 million, setting the stage for this significant transformation.
Stargate Project and Broader Partnership
This investment connects to the larger $500 billion Stargate project, a joint initiative between OpenAI, Oracle, and SoftBank announced in September. The project aims to establish five U.S. AI data centers as part of a nationwide advanced AI network.
The Ohio initiative forms part of a broader joint venture unveiled at the White House in January, where SoftBank committed $18 billion toward advancing AI infrastructure.
Modular Data Center Production
The Ohio factory will specialize in manufacturing modular data centers – portable, pre-assembled units that enable faster deployment and scalable on-site capacity. The facility will feature a small operational data center as a demonstration model.
Production of these modular data center units is scheduled to commence early next year, according to the report.
OpenAI’s Ambitious Computing Goals
During an October livestream, OpenAI CEO Sam Altman revealed the company’s staggering target: building 30 gigawatts of computing capacity at an estimated cost of $1.4 trillion. The ultimate goal involves scaling to approximately 1 gigawatt per week, despite current costs exceeding $40 billion per gigawatt.
Unlike competitors Meta and Google, OpenAI lacks advertising or cloud services revenue streams to offset the enormous expenses of constructing these data centers.
Neither SoftBank nor OpenAI immediately responded to Reuters’ requests for comment on the reported developments.
