McKinney Just Won a $919 Million AI Manufacturing Campus. The Hard Part Begins When the Machines Connect

In Brief

• The City of McKinney announced on July 13, 2026 that LITEON Technology Corp. selected the city for a North American advanced manufacturing campus — a $919 million investment, more than 650,000 square feet, and at least 600 jobs over several years.
• The campus is a conversion, not a groundbreaking. LITEON is acquiring existing industrial space at the Core5 Logistics Center and hiring in phases as renovations complete. That single fact changes the entire technology problem.
• Construction creates capacity. Whether that capacity produces value depends on decisions made in the commissioning window — the period when access is granted fastest and governed least.

Executive Summary

Announcements of this size are read as construction stories. The more useful reading is that a large, complex production environment is about to be assembled inside a building that was designed for something else, by dozens of vendors working simultaneously, over a multi-year phased schedule.

Every one of those conditions works against control. A converted building inherits pathways, power distribution, and physical zoning laid out for logistics rather than for a production process. A phased schedule means the architecture established in phase one becomes the template phase four inherits, mistakes included. And simultaneous vendor mobilization means machine builders, controls integrators, and systems implementers all requesting remote access at once, from a project team whose incentive is schedule.

None of this is a warning about attackers. It is an operations problem first. A plant that cannot be controlled cannot be run at rate, cannot prove what it produced, and cannot be restored quickly when something stops. Security is a subset of controllability, not a separate program bolted on after the ribbon is cut. The decisions that determine all of this cost very little when made early and become disproportionately expensive afterward. That asymmetry is the whole argument.

The Direct Answer

What actually determines whether a new manufacturing campus delivers its expected value?

Not the building, and not the equipment list. Both are procurement problems with well-understood answers. What determines operating value is whether, on the day the first production order runs, the organization can answer six questions: who is on the network and why, what each machine is allowed to talk to, which suppliers hold access and until when, where production and quality data lives and who can change it, how a change reaches the floor, and how fast a stopped line comes back.

Those answers are not discovered during operations. They are designed during commissioning — or not designed at all, in which case they get improvised under schedule pressure and inherited permanently. This matters more for a converted campus because conversions carry a hidden assumption: that the building is already "done." The shell exists, the roof is on, the power is in. That assumption produces a technology plan that starts when the equipment arrives instead of when the drawings are being marked up.

What Was Announced, and What Has Not Happened Yet

The verifiable facts are worth separating from the anticipation around them.

LITEON Technology Corp., a Taiwan-based manufacturer of power management and AI infrastructure products, has selected McKinney for a North American advanced manufacturing campus. The announced investment is $919 million, the campus comprises more than 650,000 square feet of existing industrial space at the Core5 Logistics Center, and the city expects at least 600 jobs over the next several years. The Office of the Texas Governor announced the same day that the initial phase represents a $307 million capital investment and 500 jobs, supported by a $3.5 million Texas Enterprise Fund grant and a $100,000 Veteran Created Job Bonus, and described the site as a North American headquarters for advanced manufacturing producing AI power and infrastructure-related solutions.

What has not happened: renovations are not complete, hiring has not begun, and nothing is in production. The city's own language is that positions open in phases as renovations are completed. Reporting on the company's application to the Texas Comptroller indicates the site is anticipated to be fully operational by 2030, with final assembly, testing and quality verification, and packaging performed at the campus alongside research and development.

That timeline is the story. A four-year phased build-out is not one commissioning event. It is a sequence of them, each one constrained by what the previous phase established.

Why a Conversion Is a Different Technology Problem

A logistics building is optimized for moving pallets: wide clear spans, dock doors along one face, power distributed for lighting and material handling, and cable pathways run where they were convenient rather than where a process would want them.

Advanced electronics manufacturing wants something else: power density and quality at specific locations, controlled environments in defined zones, network pathways that follow process flow so a cell's controllers, safety systems, and test stations share a physical and logical neighborhood, and separation between production, engineering space, and general circulation. In a purpose-built plant, process layout and infrastructure layout are designed together. In a conversion, one of them already exists.

Pathway and zoning decisions therefore get made during renovation design — months before anyone is thinking about network architecture — and they quietly determine what segmentation is achievable later. Segmentation is usually described as a configuration choice. In a converted building it is substantially a construction choice, because if the conduit does not go where the boundary needs to be, the boundary becomes a compromise. This is the same dynamic Metro Relay has written about in commercial buildings, where an inspection-ready building and a ready infrastructure are different things — it is simply more expensive on a factory floor.

Carrier and connectivity lead times deserve the same early attention. A new industrial park may not have the diverse fiber entry a production campus needs, and provisioning that capacity is a construction-schedule item, not an IT procurement item.

The Commissioning Window

Here is the observation we consider most useful, and the one most often missed.

The period between equipment delivery and stable production is the highest-access, lowest-governance moment in a plant's entire life. Machine builders need to commission their equipment. Controls integrators need to tune it. Test-equipment vendors need to calibrate and validate stations. Systems implementers need to connect the manufacturing execution layer to the enterprise. Every one of them, reasonably, asks for remote access so their specialists do not have to fly in for every adjustment. Every one of those requests arrives while the project is behind schedule, because projects of this scale are always behind schedule.

Access granted under those conditions is fast, poorly documented, and persistent — often lasting the operating life of the equipment, because nobody revisits it once the line is running and revisiting it risks a support agreement.

Machine identities compound it. They are usually shared rather than individual: a single account for "the integrator," credentials embedded in controller programs, certificates with expiry dates nobody tracks. When the integrator's contract ends, their people are offboarded from your identity system while their machine's embedded credential is not, because it lives inside a program rather than a directory.

The fix is not complicated, but it has to be a condition of the contract rather than a request after the fact: named individual access, time-bound to the commissioning scope, brokered through a controlled path rather than a vendor's own tunnel, logged, and with an expiry that is enforced rather than aspirational. Written into a purchase order, this is a paragraph. Negotiated after a machine is installed and running, it is a change order with leverage on the wrong side of the table.

What Must Be True Before First Production

Where Converted Plants Get Caught

"The equipment vendor handles security for their machine." Vendors are responsible for their product, not for your plant. The boundary between one integrator's cell and the rest of the network belongs to you, and no supplier's scope includes what happens when their equipment is reachable from somewhere it should not be.

"We will segment the network after we stabilize production." Stabilization never produces a window. Every subsequent phase adds equipment that assumes the current flat topology, so the retrofit gets larger each quarter it is deferred.

"Test stations just need patching like everything else." This is where general IT instinct fails on a factory floor. Test and measurement equipment is frequently validated as a unit — instrument, controller, and software together — with the controller running whatever operating system the instrument vendor qualified. Patching can invalidate that qualification, which is a quality problem rather than an inconvenience. The workable answer is isolation, strict control of what may communicate with the station, and monitoring.

"Backups cover us." Backups of file servers do. The question is whether controller programs, HMI configurations, robot teach points, station calibration profiles, and the manufacturing execution database can be restored to a consistent point in time. Those artifacts live in different places, are owned by different parties, and are almost never tested together.

"A spare unit means we can recover fast." A spare controller without the current program loaded and verified is inventory, not a spare. The recovery time is however long it takes to find the right program version and prove it is the right one.

The Metro Relay Connected Factory Control Model

This is Metro Relay's operational framework for sequencing control decisions in a new or converted production environment. It is not a standard, a certification, or a compliance scheme, and it does not guarantee any security or regulatory outcome. It is a way to decide what has to be settled before the next phase energizes.

Two layers carry disproportionate weight. Suppliers is where most converted plants lose control, because access decisions are made by people optimizing for schedule. Evidence is where the value is either preserved or lost, because a commissioning team disperses and takes its undocumented knowledge with it.

The Operating Consequences

A plant that cannot restore quickly does not have a security problem; it has a throughput problem denominated in shifts. A tested recovery runbook looks trivial in retrospect and expensive in advance, and that ordering is why it gets deferred.

Production and quality data is the quieter dependency. In electronics manufacturing, the test record demonstrates a unit met specification. If that record is incomplete, unattributable, or alterable, the ability to defend a field failure, satisfy a customer audit, or contain a suspect lot degrades sharply. Traceability is not an IT feature; it is the evidence supporting the warranty.

Physical access belongs in the same conversation. A logistics building is designed to move goods efficiently, which is not a design that separates production floor, engineering space, and network rooms into controlled zones. Badge boundaries follow walls, and walls are decided during renovation.

Leadership Considerations

The organizational failure mode here is a sequencing error rather than a capability gap. Technology planning typically enters a facility project when equipment is specified. By then, pathways, power distribution, zoning, and carrier entry are largely fixed, and the technology team is optimizing inside constraints it was not present to shape.

The correction is unglamorous: put someone accountable for production technology architecture into the renovation design process, before the equipment list is final, and give them authority to require conditions in supplier contracts. Their most valuable contributions will be a handful of drawings and a paragraph in a purchase order.

There is also a phase-boundary discipline worth adopting explicitly. At the end of each phase, before the next one energizes, confirm three things: that access granted during that phase has expired or been renewed deliberately, that recovery has been tested against the equipment actually installed, and that the commissioning documentation is in the operating organization's hands rather than the project team's. Phased build-outs fail slowly, one undocumented decision at a time.

Metro Relay's Perspective

Our view is that the highest-leverage window in a manufacturing project is also the shortest and the least supervised. Once a line produces revenue, the appetite for touching it drops to nearly zero, and every control that was not established beforehand becomes a negotiation with production schedules.

We also think the framing of "manufacturing cybersecurity" understates the problem in a way that costs manufacturers money. Executives hear a security program and reasonably ask what the threat justifies. The better question is whether the plant is controllable — whether you can see it, change it deliberately, restore it, and prove what it did. Those capabilities pay for themselves in uptime and audit outcomes regardless of whether anyone ever attacks the facility. Security is what you get for free when the plant is controllable.

Sequencing: What to Settle Before the Next Phase Energizes

Get technology architecture into renovation design. Pathways, zone boundaries, network room locations, power for control systems, and carrier entry are construction decisions with permanent consequences. Structured cabling and pathway design is the earliest, cheapest lever available.

Write access terms into procurement. Named individuals, brokered connections, defined scope, enforced expiry, and logging — as conditions of purchase for every machine builder, integrator, and systems supplier, before equipment ships.

Design segmentation from the process, then confirm the building can support it. Group by production cell and function, decide what must cross each boundary, and verify the physical infrastructure allows the boundary to exist where the design puts it.

Establish the change path before the first program change. Who may load a controller, what approval is required, where the authoritative program version lives, and how to roll back.

Test recovery against real equipment, not a plan. Restore a controller, an HMI, a test station, and the production database together, and time it. Business continuity in a plant means a working production state, not recovered files.

Close each phase deliberately. Expire access, retest recovery, transfer documentation, and record what was verified — then energize the next phase.

Future Outlook

North Texas is accumulating this class of facility quickly, and the regional consequence is a supplier and workforce market that will be asked to support connected production at a scale the metroplex has not previously carried. Manufacturers in Collin County and across the region should expect the questions to travel down the supply chain: customers of a campus like this increasingly ask their own suppliers to demonstrate access control, traceability, and recovery capability as a condition of the relationship.

That is the durable point beyond this announcement. Whether or not any particular project meets its schedule, the operating standard for connected manufacturing in this region is being set now, by the plants currently being commissioned.

Conclusion

A $919 million commitment to McKinney is a genuine economic event, and the city and state competed successfully for it. What that investment ultimately produces depends on a set of decisions that are inexpensive today and expensive in eighteen months: where the pathways run, who holds access and until when, what each machine may talk to, where the production record lives, and how quickly a stopped line comes back.

Buildings are commissioned once. The control architecture established while they are commissioned is inherited for as long as the plant runs.

Key Takeaways

• LITEON's announced McKinney campus is $919 million, more than 650,000 square feet, and at least 600 jobs — and it is a conversion of existing industrial space, phased over several years, not a groundbreaking.
• Converting a logistics building means pathway, power, and zoning constraints are inherited, so segmentation becomes partly a construction decision rather than purely a configuration one.
• The commissioning window is the highest-access, lowest-governance period in a plant's life, and access granted then tends to persist for the equipment's operating life.
• Supplier access terms belong in the purchase order, where they cost a paragraph, rather than in a post-installation change order.
• Test and measurement stations are frequently validated as a unit and cannot simply be patched; isolation, strict communication control, and monitoring are the workable answers.
• Recovery means restoring controllers, HMIs, teach points, calibration profiles, and the production database to a consistent state — not restoring files.
• In a phased build-out, phase one's architecture becomes the template every later phase inherits, which is why each phase should close deliberately before the next energizes.

Frequently Asked Questions

Is the LITEON campus operating yet? No. As announced in July 2026, LITEON has selected McKinney and is acquiring existing industrial space. Renovations are underway or planned, hiring is expected to begin as renovations complete, and positions are described as opening in phases over several years. Reporting on the company's filing with the Texas Comptroller indicates the site is anticipated to be fully operational by 2030.

Why does converting an existing building matter technically? Because the building's pathways, power distribution, and physical zoning were designed for a different use. Those constraints are fixed during renovation and then limit what network segmentation and physical access separation are achievable afterward.

When should technology planning start on a facility project? During renovation or design development, well before equipment is specified. Conduit routes, network room locations, control-system power, and carrier entry are construction decisions, and they are the cheapest lever available.

What is the practical risk of vendor remote access? It is usually granted quickly during commissioning, documented poorly, and never revisited. The compounding issue is that machine credentials are often shared or embedded in controller programs, so ending a vendor relationship does not necessarily end the access.

Why can't production equipment just be patched on a normal schedule? Much of it is qualified as a validated combination of instrument, controller, and software. Changing the controller's operating state can invalidate that qualification, creating a quality problem. Isolation and strict control of permitted communication are generally the workable approach.

What does recovery actually mean in a plant? Returning to a working production state — controller programs, HMI configurations, robot teach points, calibration profiles, and the manufacturing execution database restorable together to a consistent point in time. That is a materially harder test than restoring a file server.

Does this only apply to campuses of this size? No. The sequencing applies to any manufacturer adding a line, converting space, or bringing connected equipment onto the floor. The scale changes the cost of getting it wrong, not the order of the decisions.

Manufacturing Technology Readiness Review

Metro Relay works with manufacturers and multi-site operators across Dallas–Fort Worth on the technology decisions that have to be made before production starts: pathway and segmentation design, supplier access terms, machine identity, production data handling, change control, and tested recovery.

A Manufacturing Technology Readiness Review examines a planned or in-progress facility against the Connected Factory Control Model and identifies which decisions are still inexpensive — and which are about to stop being. If you are converting space, adding a line, or entering a commissioning phase in Collin County or anywhere across North Texas, start the conversation, or review Metro Relay's cybersecurity and structured cabling capabilities.

Sources

• City of McKinney, McKinney Welcomes $919 Million LITEON Investment, More Than 600 New Jobs, July 13, 2026. Supports: LITEON's selection of McKinney; the $919 million announced investment; more than 650,000 square feet of existing industrial space at the Core5 Logistics Center; at least 600 anticipated jobs; LITEON's description as a power management and AI infrastructure company; and phased hiring as renovations are completed. https://www.mckinneytexas.org/CivicAlerts.aspx?AID=6930
• Office of the Texas Governor, Governor Abbott Announces LITEON Expansion in McKinney, July 13, 2026. Supports: the description of the site as a North American headquarters for an advanced manufacturing facility producing AI power and infrastructure-related solutions; the initial phase figure of $307 million and 500 jobs; and the $3.5 million Texas Enterprise Fund grant and $100,000 Veteran Created Job Bonus. https://gov.texas.gov/news/post/governor-abbott-announces-liteon-expansion-in-mckinney
• Manufacturing Dive, Lite-On Technology to establish $919M manufacturing campus in Texas, July 2026. Reputable secondary reporting used where the underlying filing is not publicly accessible. Supports: the campus scope including research and development; final assembly, testing and quality verification, and packaging at the site; and the anticipated fully operational date of 2030 drawn from the company's application to the Texas Comptroller of Public Accounts. https://www.manufacturingdive.com/news/lite-on-electronics-r-d-manufacturing-919-million-mckinney-texas/825250/