Foxconn and OpenAI Discuss New AI Hardware Deal as US Server Production Expands

Foxconn is preparing a new collaboration with OpenAI as the Taiwanese manufacturing giant expands its role in the global AI infrastructure industry. Chairman Young Liu confirmed talks with OpenAI CEO Sam Altman, but the companies have not disclosed the scope, financial terms or timing of a potential agreement.

Liu met Altman on Friday, October 30, before speaking to reporters at Foxconn's annual technology event in Taipei. According to Reuters, Foxconn plans to announce further details once the arrangements are finalized.

The discussions come as Foxconn expands its American manufacturing operations and AI companies increasingly depend on specialized server hardware to support larger computing clusters.

The companies are building on a partnership announced in November 2025, when they agreed to collaborate on the design and engineering of next-generation AI data center equipment.

Under the original agreement, OpenAI would provide insights into the infrastructure requirements of future AI systems. Foxconn would use that information to develop hardware suitable for production at its US manufacturing facilities.

The collaboration covers several generations of AI server racks, alongside critical components such as networking equipment, cabling, cooling systems and power infrastructure.

One important distinction is that the 2025 agreement did not include binding purchase commitments or financial obligations.

Instead, OpenAI received early access to evaluate the resulting hardware and an option to purchase it.

That structure allowed both companies to explore manufacturing requirements without immediately committing to large equipment orders.

The latest discussions could expand those arrangements, but neither company has confirmed whether they involve additional manufacturing capacity, new products or binding orders.

Foxconn, formally known as Hon Hai Precision Industry, is best known for assembling Apple's iPhones. Increasingly, however, AI servers are becoming central to its business.

The company is Nvidia's largest AI server manufacturing partner, according to Reuters. It operates six factory buildings in Houston, Texas, dedicated to manufacturing AI servers for the chipmaker.

Liu said further expansion of those operations would be necessary.

The shift toward AI infrastructure is already visible in Foxconn's financial performance.

During the second quarter of 2026, the company's cloud and networking products division accounted for 51% of total revenue, exceeding half of the business for the first time.

The division includes AI servers, although Foxconn does not attribute the entire segment's revenue to AI.

In its second-quarter financial results, Foxconn reported total revenue of NT$2.53 trillion and highlighted continued demand for AI infrastructure.

The company's growing manufacturing footprint gives it an important position between chip designers, component suppliers and the companies building large AI computing facilities.

That position also makes Foxconn relevant to investors tracking AI infrastructure beyond Nvidia, AMD and the major cloud providers.

Nvidia's GPUs are among the most valuable components inside modern AI systems, but the processors alone cannot operate an AI data center.

Large computing clusters require tightly integrated racks containing accelerators, processors, high-speed networking equipment, power distribution systems and increasingly sophisticated cooling technology.

A single unavailable component can delay the installation of an entire system.

Manufacturing these systems involves coordinating multiple suppliers, assembling complex hardware and testing equipment before it can be installed in a data center.

As described in AI World Today's guide to Nvidia's global chip supply chain, semiconductor fabrication represents only one part of the process. Advanced packaging, memory production, server integration and final assembly are also essential.

Foxconn operates near the end of that chain, turning sophisticated electronic components into systems that data center operators can deploy.

The original OpenAI partnership also emphasized a broader US manufacturing ecosystem.

Rather than relying exclusively on existing supplier arrangements, the companies planned to improve rack designs, expand domestic sourcing options and increase localized testing and assembly.

For OpenAI, greater involvement in hardware design could help align equipment with the requirements of future models.

For Foxconn, a closer relationship with a leading AI developer could provide earlier insight into demand for upcoming generations of computing systems.

These are potential strategic advantages, not confirmed financial benefits from the proposed new agreement.

The potential expansion of Foxconn's OpenAI relationship comes during a period of heightened concern about AI infrastructure spending.

Technology companies are committing enormous amounts of capital to data centers, computing equipment and electricity infrastructure. Investors are increasingly questioning whether future AI revenue will justify those costs.

Foxconn's chairman remains optimistic.

Speaking at the Taipei event, Liu rejected the suggestion that the AI industry is experiencing a speculative bubble.

Foxconn expects double-digit revenue growth in 2027, supported by demand for AI-related manufacturing capacity.

That forecast reflects management's expectations rather than guaranteed future performance.

The distinction between manufacturing demand and financial commitments is particularly important.

A hardware development partnership does not automatically translate into equipment sales. Likewise, plans for additional data center capacity do not necessarily establish when the hardware will be ordered, delivered or recognized as revenue.

The broader AI infrastructure market faces similar challenges involving equipment availability, construction schedules, power access and financing.

Foxconn's position as a major Nvidia server manufacturer provides exposure to expanding AI computing demand. At the same time, its growth depends on customers following through with large infrastructure projects.

For now, the next milestone is a formal announcement from Foxconn and OpenAI.

That announcement will need to clarify whether the companies are extending their existing design collaboration or committing to new production volumes and equipment purchases. Until then, the value of any additional agreement remains unknown.