Sources: SoftBank plans to invest up to $3B to remodel an EV plant in Lordstown, Ohio, that will produce equipment for OpenAI's forthcoming US data centers
Sri Muppidi / The Information :
Context & Ripple Effects
SoftBank had already acquired the Ohio facility in a $375M purchase aimed at Stargate; a subsequent account said Foxconn would continue operating it and make AI servers. The reported retrofit would turn that ownership into a dedicated manufacturing commitment for OpenAI-linked data-center buildouts.
The move fits SoftBank’s broader reuse of industrial sites for AI infrastructure, including its reported plan to convert a former Sharp display factory in Osaka into a data center for work with OpenAI.
First-order effects
- SoftBank would commit up to $3B to retool the Lordstown EV plant for equipment used in OpenAI’s forthcoming U.S. data centers, shifting the site’s output from electric-vehicle manufacturing toward AI-infrastructure supply.
- OpenAI gains a prospective, SoftBank-backed domestic equipment source for its planned facilities, while Foxconn’s continuing plant role becomes more closely tied to AI-server and related infrastructure production.
Second-order effects
- The retrofit connects factory capacity directly to the Stargate buildout, increasing the importance of execution across equipment production, site construction, and power availability rather than treating data-center expansion as a standalone real-estate project.
- Ohio’s AI-infrastructure cluster would require coordinated investment beyond the plant itself; SoftBank’s later-reported 9.2GW power-plant consortium plan illustrates how electricity supply is becoming coupled to the same buildout.
Third-order effects
- If repeated, the pattern would make idled or repurposable industrial assets a strategic input to AI capacity, with financiers and frontier labs assembling manufacturing, energy, and data-center infrastructure in one investment chain.
- It also concentrates buildout risk: delays at any link—equipment, power, or data-center development—could constrain the usable compute capacity that the broader plan is intended to create.
The trend: AI infrastructure is evolving from data-center leasing into vertically coordinated financing of factories, power, and compute supply for a small number of frontier labs.