Intel Fab 62 is the center of Intel’s next bet on leading-edge manufacturing. The company has begun installing production tools at its sixth high-volume building at the Ocotillo campus in Chandler, Arizona, targeting the start of operations by the end of 2028. Each ASML High-NA EUV scanner costs roughly $380 million, about double a standard extreme ultraviolet machine.
The installation is significant because a completed factory shell is not yet a working semiconductor fab. Tool move-in marks the transition from construction to factory integration, bringing the equipment, utilities and process controls needed to turn silicon into commercial chips. Intel has completed the building shell and is now positioning production equipment for the planned start of operations.
Intel Fab 62 brings a High-NA EUV bet into Arizona
Intel’s strategy at Fab 62 is closely connected to ASML’s High-NA EUV technology. Each scanner costs roughly $380 million, about double the price of a standard EUV system, and arrives in approximately 250 shipping crates with a combined weight of 330,000 pounds. ASML and Intel engineers need several months to install and calibrate the machinery before production wafers can pass through it.
High-NA changes the economics and capabilities of leading-edge lithography. Standard EUV scanners use a numerical aperture of 0.33, while High-NA systems raise it to 0.55. The redesigned anamorphic optics can print features that would otherwise require several overlapping exposures, reducing the need for repeated masking steps. ASML’s Twinscan EXE:5200B can process more than 175 wafers per hour with 0.7-nanometer overlay accuracy.
Intel is not waiting for the technology to mature in a distant production cycle. It installed the first commercial High-NA scanner at its D1X development facility in Oregon in April 2024, and production layers were printed during the summer of 2026. Selected 18A layers for the Panther Lake family were manufactured with the system. Intel and ASML have also reported processing more than one million wafers across development, certification and production work.
Intel has committed to using High-NA for its 14A node. TSMC, which controls about 62 percent of global contract foundry revenue, has postponed commercial adoption until 2030 and continues using standard EUV with multiple patterning passes. That creates an opportunity for Intel to gain an early-production advantage, but only if the newer equipment produces acceptable yields and enough tools arrive on schedule.
Tool move-in is the decisive stage of a fab project
The scale of the equipment explains why installation at Fab 62 matters more than a routine construction update. A leading-edge fab needs lithography systems, deposition reactors, plasma etchers, implanters, polishing equipment, metrology tools and a tightly controlled flow of chemicals and gases. The complete tool package can cost $10 billion or more before the full cost of buildings, utilities and workforce is counted.
Intel has said Fab 42, Fab 52 and Fab 62 represent about $32 billion in combined investment. Fab 52 and Fab 62 alone are estimated at roughly $25 billion for their combined tool and infrastructure buildout. Those figures show why tool delivery and factory utilization matter. Contract manufacturers operate with high fixed costs, so a technically advanced factory must sustain high cleanroom utilization to generate acceptable returns.
Fab 52 provides a useful model because it is already producing 18A chips at high volume. Fab 62 could support more 18A output, the enhanced 18A-P process with backside power delivery, or Intel 14A. Intel has not publicly fixed the node mix or disclosed planned wafer capacity, leaving room to adjust the factory as customer demand develops.
That flexibility can reduce risk. Intel designed 14A to reuse about 80 percent of the manufacturing equipment associated with 18A, which could moderate the capital required for the next node. Fab 62 can therefore function as flexible capacity rather than a one-purpose factory tied to a single customer or process.
14A needs outside customers to justify the next spending wave
Intel has said it will not commit all the capital needed for 14A high-volume manufacturing without meaningful external foundry commitments. Risk production for internal products remains targeted for the second half of 2027, with high-volume manufacturing planned for 2028. The process design kit has progressed toward its customer-release stage, and potential interest has included major names such as Apple, NVIDIA, Google and AMD.
A signed customer would strengthen the case for Fab 62 as a 14A production center. It would also provide evidence that Intel can draw outside business despite the scale advantage of established competitors. TSMC’s much larger capital budget, supply-chain purchasing power and broad customer relationships cannot be overcome by lithography alone.
Intel nevertheless has advantages. Its High-NA systems have already worked on commercial logic layers, while TSMC has delayed volume deployment. Backside power delivery may help justify the high cost of the equipment by freeing frontside routing space and reducing the need for complicated multi-patterning on some layers. High-NA does not automatically deliver a smaller node, but it can give Intel a more advanced manufacturing toolkit sooner.
Production risk remains substantial. The smaller exposure field of High-NA scanners requires mask stitching for large chips, including artificial intelligence accelerators, and stitch defects can reduce yields. Intel has addressed part of the problem by qualifying important layers for both standard and High-NA exposure. True benefits from finer High-NA geometry will become clearer when 14A reaches volume production.
Fab 62 cannot erase the wider schedule pressure
Intel’s Ohio One project illustrates the difference between a firm tool plan and a large project still exposed to changing schedules. Workers installed the first major structural truss at the first New Albany fab, marking the start of above-ground construction. Intel originally expected production in 2025, later moved it to 2027 and 2028, and then pushed construction completion to 2030 with production in 2030 or 2031. Updated milestones place the first Ohio fab’s production as late as 2032.
The delay threatens more than the schedule. Ohio state grants of $300 million per factory were tied to construction completion by the end of 2028. The revised timeline means Intel could lose access to that support, although the project remains eligible for at least $1.5 billion in CHIPS Act funding. The scale is substantial, with two planned fabs, about 2.5 million square feet of buildings and 600,000 square feet of cleanroom production space in the first phase.
Intel’s Arizona schedule is tighter, but the market’s reaction will depend on execution and demand. Tool delivery, installation, process qualification and customer orders all have to align before Fab 62 changes Intel’s financial trajectory. Early production will not guarantee immediate foundry leadership, and a new leading-edge fab can carry years of depreciation before it reaches optimal output.
A concrete test for Intel’s foundry strategy
Intel Fab 62 is now a concrete test of whether advanced equipment can become competitive volume manufacturing. Its High-NA tool move-in gives Intel a possible lead over TSMC in commercial 0.55-NA production, while the planned 2028 start creates a narrow opportunity to establish 14A credibility before rival adoption accelerates. The benefits will depend on yield, uptime, tool supply, customer commitments and factory utilization rather than scanner specifications alone.
For investors and customers, the important signals will be the final equipment mix, confirmed node assignments, production timing and named external orders. Intel Fab 62 may ultimately combine 18A capacity with 14A production, but the flexible strategy itself is rational. It lets the company respond to demand while pressing ahead with the process and lithography needed to challenge the market leader.
The next phase will test whether Intel can execute at the scale implied by its ambitions. With Fab 62 tools entering the building, Intel Fab 62 has moved beyond an announced future and into the difficult work of building a competitive foundry operation from the equipment up.