TAIPEI, Taiwan (AP) — OpenAI and Taiwan electronics giant Foxconn have agreed to a partnership to design and manufacture key equipment for artificial intelligence data centers in the United States as part of ambitious plans to fortify American AI infrastructure.

Foxconn, which makes AI servers for Nvidia and assembles Apple products, including the iPhone, will be co-designing and developing AI data center racks with OpenAI under the agreement, the companies said in separate statements late last week.

The products Foxconn will manufacture in its American facilities include cabling, networking and power systems for AI data centers, the companies said. OpenAI will have “early access” to evaluate and potentially to purchase them.

Foxconn has factories in the U.S., including in Ohio, Wisconsin and Texas. The initial agreement does not include financial obligations or purchase commitments, the statements said.

The Taiwan contract manufacturer, formally known as Hon Hai Precision Industry Co., has been moving to diversify its business, developing electric vehicles and acquiring other electronics companies to build out its product offerings.

A Model A EV made by the group’s automaking affiliate Foxtron was on display at an event last week.

“This year, Model A. ‘A’,’ for affordable,” said Jun Seki, chief strategy officer for Foxconn’s EV business.

The tie-up with OpenAI can also help Taiwan, a self-governed island claimed by China, to build up its own computing resources, said Alexis Bjorlin, a Nvidia vice president.

“This allows Taiwan’s domain knowledge and key technology data to remain local and ensure data security,” she said.

“This partnership is a step toward ensuring the core technologies of the AI era are built here,” Sam Altman, CEO of San Francisco-based OpenAI, said in the statement. “We believe this work will strengthen U.S. leadership and help ensure the benefits of AI are widely shared.”

OpenAI has committed $1.4 trillion to building AI infrastructure. It recently entered into multi-billion partnerships with Nvidia and AMD to expand the extensive computing power needed to support its AI models and services. It is also partnering with American chipmaker Broadcom in designing and making its own AI chips.

Its massive spending plans, however, have worried investors, raising questions over its ability to recoup its investments and remain profitable.

Altman said this month that OpenAI, a startup founded in 2015 and maker of ChatGPT, is expected to reach more than $20 billion in annualized revenue this year, growing to “hundreds of billions by 2030.”