Intel Fab Buildout: Arizona Tool Install Starts, Ohio Sets Truss

What Intel announced

Intel's long-stalled U.S. factory buildout is moving again, and the most important evidence is in Arizona. Workers have started installing manufacturing equipment at Fab 62 on the company's Ocotillo campus near Phoenix. In New Albany, Ohio, crews have raised the first major truss at the site's first fab. Naga Chandrasekaran, who leads Intel Foundry's technology and operations, disclosed both milestones in a company post on October 21, 2026, and Bloomberg was first to report it.2122 He described the Ohio work as another significant step toward more leading-edge chip capacity in the U.S.2

The two milestones should not be weighed equally. Fab 62's shell was finished some time ago, so tool installation means the project has moved past building construction and into its final, most expensive phase.27 A roof truss in Ohio is a real construction marker, but it does not show that Intel has decided to equip that fab and run it. Most coverage put both items under one "expansion resumes" headline. In practice, Arizona is the capital commitment and Ohio is the signal.

Arizona: the expensive phase begins

Fab 62 sits next to Fab 52, which is already producing chips. It will be the sixth high-volume fab at Ocotillo.2227 Equipping a fab costs more than any other stage of getting it ready for production, because single tools can cost hundreds of millions of dollars.25 Starting that work is a bigger financial decision than putting up a building.

The reports disagree on timing. Bloomberg-based accounts, including the Los Angeles Times and The Next Web, say Intel expects Fab 62 to start operating in early 2028.2529 CNBC, Tom's Hardware and Wccftech put operations by the end of 2028.22722 The difference may simply reflect how loosely each outlet described the same timeline, but nearly a year separates the two readings. Readers should treat "some time in 2028" as the firm part.

Intel has also not said what Fab 62 will make. When it was first announced, Fab 62 was meant to build Intel's most advanced processes alongside Fab 52. Fab 52 has since been assigned to 18A, while Fab 62 still has no assigned process.27 Tom's Hardware reports that the plant could run 18A, 18A-P or 14A. Intel has not given a wafer capacity or a capital cost for the facility.27 The 14A option matters most. Intel has said 14A should enter risk production in the second half of 2027 and reach high volume in 2028, which lines up closely with Fab 62's schedule.17 That is an inference, not a disclosed plan, but it is the most obvious reason to equip the building now.

Ohio: progress on a much-delayed project

The New Albany project is a different story. When Intel announced it in 2022, it planned two leading-edge fabs, investment of more than $28 billion, and first chips in 2025.24 In February 2025, Intel moved the first module's construction finish to 2030, with operations in 2030 or 2031. The second module is now due to finish in 2031 and start operating in 2032.1 That puts the site at least five years behind its original schedule.2

Coverage splits on whether this week's update changes that schedule. CNBC, Stocktwits and BigGo repeat the 2030 completion target.22423 Tom's Hardware points out that Intel gave no updated schedule, no process node and no capacity figure for Ohio.27 The more careful reading is that nothing has changed. A truss confirms the building is going up, but nothing Intel said brings the production dates forward.

Ohio has absorbed real money anyway. Intel reported spending $1.53 billion there in 2025, bringing total capital spending at the campus to $5.26 billion. That is a lot of work, but it is less than a fifth of the $28 billion initial commitment.5 Intel's own February 2025 update explained the approach: build more slowly, but keep the option to speed up construction and the start of production if customer demand supports it.1 The truss fits that slower pace. It does not show a switch to a faster one.

Why the tone has changed

The reason for the more upbeat messaging is mostly Intel's own product demand, not foundry customers. Intel is reportedly struggling to keep up with orders for data center processors used for AI services.25 In the second quarter of 2026, revenue rose 25% from a year earlier to $16.1 billion, and data center and AI revenue grew 59%.16 Intel also raised its 2026 capital spending forecast from $18 billion to more than $20 billion and said 2027 would be higher.16 CFO David Zinsner said most of the extra spending would go to front-end fab capacity rather than packaging, mainly in Intel's U.S. network.18

That background explains why this matters. Since taking over in March 2025, CEO Lip-Bu Tan has said big projects such as Ohio would go ahead only when there was enough work to fill them.21 Equipping a new Arizona fab is the clearest sign yet that Intel believes the demand case for that site has been met. The Investing.com summary adds that Intel has not announced any broader change in spending policy.21 Both things can be true. Intel is still spending in line with demand, and its demand estimates have gone up.

The foundry customer problem

The weak point is external customers. In the second quarter, Intel Foundry booked $5.8 billion in revenue, but only $293 million came from outside customers, about 5% of the total. The segment lost about $2.1 billion.1612 Intel's FY2025 annual report says the company has few external foundry customers so far and names none. It also added a new risk factor warning that 14A could be paused if Intel does not win a significant external customer.19 So the equipment in Fab 62 is a bet that sits next to a disclosed risk the next node might not go ahead.

There are some encouraging signs, but none have been confirmed. In September, reports said SK Hynix had discussed leasing part of the Ohio site to make memory chips. Neither company has confirmed a plan.223 Apple was reported in May to be in early talks about using Intel's factories, and Intel has joined Elon Musk's Terafab project as a manufacturing partner.29 TechInsights analyst Dan Hutcheson said major chip designers want a second source to put pricing pressure on TSMC and have been running tests at Intel.23 Still, testing a process is not the same as signing a volume contract. Until a named customer commits to 14A or to Ohio, these new fabs depend mainly on demand for Intel's own products.

The spending gap

Intel's higher budget still leaves it far behind its rivals. Analysts expect TSMC to spend about $60 billion this year, roughly three times Intel's figure, and Samsung is also spending considerably more.2623 Intel's balance sheet limits how far it can go. It ended June with about $29.7 billion in cash and investments, raised about $23 billion more in an August share offering, and carries about $50.5 billion in debt. Its adjusted free cash flow in the second quarter was negative $8.4 billion.19 The U.S. government also holds a stake in the company, which gives Washington a direct interest in whether these fabs succeed.29

Intel does have one strength rivals lack in the U.S. Tom's Hardware notes that Intel does leading-edge process research, photomask production and advanced packaging domestically. TSMC and Samsung do not run all of those in America.27 Intel also confirmed it has put High-NA EUV lithography into high-volume manufacturing at its Oregon development fab.27 These capabilities help Intel's case as a domestic supplier, but they do not replace paying customers.

What to watch

The next test comes soon. Intel reports third-quarter results on October 29. The company guided to revenue of $15.8 billion to $16.8 billion, and analysts expect about $16.5 billion.1114 More important than the headline numbers will be any update on 2027 capital spending, any process assignment for Fab 62, and any external foundry customer. Overall, Intel has moved from a slow pause to targeted spending, putting money into a site that can produce chips in 2028 while keeping Ohio as a slower-moving option. If an outside customer is named, the Ohio truss could become the start of a faster buildout. If not, Ohio is likely to stay on its delayed schedule.