OpenAI Opens RFP to Strengthen U.S. AI Supply Chain

OpenAI has launched a Request for Proposals (RFP) aimed at U.S.-based manufacturers across the AI supply chain, with submissions due by June 2026. The company frames the move as the next step after its Stargate infrastructure initiative, which it says is already more than halfway to a 10-gigawatt capacity commitment. The RFP targets components that rarely make headlines but that OpenAI argues will decide whether America's AI build-out is durable: racks, cabling, cooling, power electronics, motors and gearboxes.

The announcement matters because it extends OpenAI's role from a buyer of data centers to a shaper of domestic manufacturing. Where public debate concentrates on chips and gigawatt campuses, the company is explicitly trying to redirect attention — and capital — toward the unglamorous physical layer underneath. Whether a single RFP can move that layer is a separate question, and OpenAI's own language leaves it open: responses will "help inform partnerships, procurement strategies, and infrastructure planning," which is coordination language, not a purchase order.

What is OpenAI actually asking for?

The RFP, published on OpenAI's site under the title "Strengthening the U.S. AI supply chain through domestic manufacturing," invites manufacturers, suppliers and partners that are building — or prepared to build — critical components and systems inside the United States. The company lists three focus areas:

• Modules, tooling and equipment, and final assembly for consumer electronics
• Manufacturing for compute, power, cooling and supporting data center hardware
• Critical inputs for advanced robotics, with OpenAI citing gearboxes, motors and power electronics as examples

The scope is deliberately broad. It spans everything from the sheet-metal and cabling inside a data center hall to the electromechanical parts of robots, a category that signals OpenAI's interest in physical AI extends beyond servers. The RFP does not specify volumes, dollar commitments or guaranteed offtake. OpenAI says it will review proposals on a rolling basis and follow up with selected respondents on next steps; no figures on anticipated procurement appear anywhere in the announcement.

That absence is the caveat worth holding onto. An RFP is a demand signal, and demand signals can unlock private investment — but nothing in the document obligates OpenAI to buy anything.

Why OpenAI says infrastructure is destiny

The blog post's argument rests on a simple premise, stated in its own words: "Infrastructure has long been destiny when it comes to America's economic success, and that will be especially true in the Intelligence Age." OpenAI contends that building advanced AI infrastructure is a "historic opportunity" to strengthen domestic supply chains and reindustrialize the country — catalyzing U.S. manufacturing, modernizing the energy grid, creating well-paid jobs and reinforcing American leadership.

The stakes the company identifies are economic and strategic at once. If the AI build-out depends on foreign-made racks, cooling systems, power electronics and assembly capacity, the resilience of the entire ecosystem sits outside U.S. borders. OpenAI's stated goal for the RFP is to "shorten timelines, strengthen resilience, and extend technology leadership as AI infrastructure scales." Those are three promises in one sentence, and each depends on manufacturers responding with capacity that does not yet exist at the required scale.

OpenAI also makes a claim about jobs that it presents as already underway. "These investments are already translating into good jobs and local economic growth in communities across the country," the company writes of its Stargate-era spending. No headcount, wage or location figures accompany the claim. Readers should treat it as an assertion by an interested party rather than an audited result.

How far along is Stargate, really?

By OpenAI's own account, Stargate is approaching its first anniversary, and the company says it has "announced planned capacity" that puts it "well over halfway" to its 10-gigawatt commitment. The operative word is "announced." Planned capacity is not delivered capacity, and the post does not break down how much of that pipeline is under construction, financed or energized.

The 10-gigawatt figure itself deserves a parenthesis for scale: one gigawatt roughly equals the output of a large nuclear plant, so the commitment implies data center campuses with the power draw of several such plants. That is precisely why OpenAI argues the conversation cannot stop at chips and data centers. In the company's telling, "the racks, cabling, networking gear, cooling systems, power systems, power electronics, electromechanical modules, and testing and assembly capacity are all required to bring it all online at scale." In plain words: the bottleneck is shifting from silicon to everything around the silicon.

OpenAI positions the RFP as building on momentum it says already exists. "Across the country, manufacturers are investing in advanced production capabilities to support the AI ecosystem – bringing new facilities online, modernizing supply chains, and expanding skilled workforces," the post states. Those investments, the company argues, "are critical to ensuring that the benefits of AI are created and shared here in the U.S." Again, no facility names, counts or timelines are given — the claim is directional, not evidentiary.

What happens next?

The mechanics are straightforward. Interested parties submit proposals, OpenAI reviews them on a rolling basis, and selected respondents hear about next steps. The deadline is June 2026 — a window long enough to admit manufacturers still raising capital for new lines, not just those with capacity in hand.

For suppliers, the significance is the demand signal itself: the largest AI developer in the U.S. is formally inviting domestic producers of motors, gearboxes, power electronics, cooling systems and assembly capacity into its planning process. For policymakers, it is evidence that industrial policy pressure around AI is now coming from the demand side of the market, not only from Washington. For competitors, it sets a marker that supply-chain positioning may become as contested as model performance.

The test now shifts from language to contracts. Until OpenAI converts rolling RFP reviews into signed procurement and named manufacturing partners, the announcement functions as intent — well-timed, credibly motivated, but still intent.