There is a building beside I-15 in Lehi that most of Utah County has driven past several thousand times without once wondering what happens inside it. It is large, pale, almost entirely windowless, and it does not advertise. It has had three owners in twenty years. It has made three genuinely different products. And it is now the site of what has been repeatedly described as the largest single economic investment in the history of the state of Utah.
Nobody local writes about it. There are guides to every canyon, every taco stand and every ward boundary in this valley, and almost nothing about the factory that put Utah County on the federal government's semiconductor map. This is that article.
It started as a joint venture, not a factory
The story does not begin with a company deciding to build a plant in Utah. It begins with two companies deciding to share the risk of one.
In January 2006, Intel and Micron Technology formed IM Flash Technologies, a joint venture headquartered in Lehi. Both partners wanted memory manufacturing capacity and neither wanted to carry the full cost of a modern fabrication plant alone — an entirely rational instinct, because a fab is one of the most capital-intensive objects a private company can build. The venture ran a 300-millimeter wafer fab in Lehi and later opened a second facility in Singapore.
That structure matters for understanding everything that came after. The Lehi building was never conceived as a permanent home for a product line. It was conceived as shared capacity, and shared capacity is the first thing that gets renegotiated when partners stop agreeing.
What IM Flash actually made
For most of its life the Lehi fab made NAND flash memory — the storage in solid-state drives, memory cards and phones — and it made it aggressively. IM Flash pushed to smaller design rules quickly, and the site had a genuine reputation inside the industry for process capability.
Later the venture's flagship product became 3D XPoint, a memory technology Intel and Micron announced jointly and pitched as sitting between conventional flash storage and system memory: faster than a drive, cheaper than RAM, and non-volatile. Intel sold its version under the Optane brand.
It was, on the engineering merits, a real achievement. Commercially it never found the volume the two companies needed it to find, and that is the pivot on which this whole story turns.
Why the partners separated
In July 2018 Micron and Intel announced they would finish the second generation of 3D XPoint together and then pursue the technology independently. Joint development of the flagship product was over.
That left an awkward asset: a large, capable, operational fab, jointly owned, making a product whose future both owners had just declined to share. In 2019 Micron bought out Intel's interest in the venture, and IM Flash Technologies formally ceased to exist as a separate company on 31 October 2019.
Micron now owned Lehi outright. The purchase of Intel's stake was reported at around $1.5 billion.
Micron bought it, then decided it did not want it
This is the part of the sequence that people find hardest to believe, and it is the part most worth understanding, because it explains why the building was available at all.
In March 2021 — roughly eighteen months after completing the buyout — Micron announced it was ending 3D XPoint production and looking for a buyer for the Lehi plant. The company was candid about the reason: the product was not generating the returns required to justify continuing.
So a company spent about $1.5 billion to take full ownership of a factory and then, before two years were out, put it on the market. Not because the factory was bad. Because the thing the factory was built to make had stopped making sense.
Texas Instruments bought a building, not a business
In June 2021 Micron announced a definitive agreement to sell the Lehi fab to Texas Instruments. Micron described the total economic value to itself as approximately $1.5 billion — about $900 million in cash from TI, with the remainder in the value of tools and other assets Micron retained, redeployed to its other sites, or sold separately. Micron confirmed the sale closed on 22 October 2021.
The $900 million figure and the $1.5 billion figure both circulate locally and people argue about which is right. Both are right; they measure different things. TI paid $900 million for the site. Micron valued the whole transaction, including equipment it kept, at around $1.5 billion.
What TI bought was an operational 300 mm building with a trained workforce, in a state it had no prior manufacturing presence in. Micron said TI would offer employment to the site's staff on closing. What TI did not buy was a memory business — it had no interest in making memory at all.
The conversion, and why it took a year
TI renamed the site LFAB, later LFAB1 once a second building was planned. Production began in December 2022, about a year after the purchase closed.
A year sounds slow for a plant that was already running. It is not. Converting a fab from one product family to another is not like retooling a car line — process recipes, equipment configurations, contamination controls and qualification testing all have to be rebuilt around the new product, and a fab does not ship anything until the output passes qualification. That the site came up in roughly twelve months is a reflection of how good the building and the workforce already were.
What the fab makes now, and why you have never heard of it
TI makes analog and embedded processing semiconductors in Lehi. These are the two categories the general public never learns the names of, and they are almost certainly the most numerous chips in your life.
An analog chip's job is to deal with the physical world. Temperature, pressure, sound, light, motion and voltage are all continuous, messy, analog quantities, and something has to convert them into numbers a digital processor can act on — and then manage the power flowing through the whole device without setting it on fire. An embedded processing chip runs one defined job inside a product: the controller in a washing machine, the module watching a battery pack, the small brain inside an industrial sensor.
A modern car contains hundreds of these. A phone contains dozens. None of them appear on a spec sheet, and none of them are what people mean when they say "chip shortage" — except that the chip shortage was, overwhelmingly, a shortage of exactly these parts.
Mature nodes are not old nodes
The federal CHIPS documentation describes the Lehi site as producing chips in the 28-nanometre to 65-nanometre range. TI's earlier statements described the acquired fab starting at 65 nm and 45 nm, with room to move beyond those nodes as required.
Anyone who follows consumer technology will notice those numbers are far from the leading edge, where the marketing runs in single digits. That comparison is a category error. Node size is a proxy for transistor density, which matters enormously for a processor whose entire value is computation per watt. It matters far less for a chip whose job is to survive under a car hood for fifteen years, tolerate voltage spikes, and be available in identical form a decade after it was designed in.
Mature-node parts are also the ones a supply chain cannot substitute its way around. That is precisely why they became a national-security conversation.
LFAB2, and the largest investment in Utah's history
In 2023 TI announced a second 300 mm fab in Lehi — LFAB2 — built next to the existing one, described as an approximately $11 billion project and widely reported as the largest single economic investment in Utah's history. TI has said the two buildings will eventually connect and operate as a single fab rather than as two neighbors, sharing infrastructure, staff and suppliers.
TI has also stated design goals for the new building: LEED Gold version 4 certification, a target of 100% renewable electricity, and water recycling at close to twice the rate of the existing fab. Those are stated intentions from the company rather than verified outcomes, and they are worth reading as such until the building is running.
The CHIPS Act money, and what it is actually contingent on
In December 2024 the U.S. Department of Commerce announced an award agreement of up to $1.6 billion in direct CHIPS and Science Act funding for TI, following a preliminary memorandum of terms announced in August 2024. The award covers three fabs: SM1 and SM2 in Sherman, Texas, and LFAB2 in Lehi.
Two details get lost every time this is summarized locally.
First, the money is milestone-based. It is disbursed against completed construction, tooling, production and commercial milestones — not paid up front. Second, the direct funding is the smaller half of the federal support. TI has also pointed to an expected investment tax credit through the Treasury Department worth several billion dollars more, tied to qualifying U.S. manufacturing investment.
Recipients also accept conditions, including restrictions on stock buybacks and national-security guardrails limiting the sharing of intellectual property. This is not a grant in the ordinary sense; it is a contract with performance terms.
The jobs number, stated honestly
Local coverage routinely garbles this, so it is worth separating.
TI's 2023 announcement put the Lehi expansion at roughly 800 semiconductor jobs. The federal documentation for the full three-fab package across Texas and Utah describes more than 2,000 manufacturing jobs and over 20,000 construction jobs over time. Those larger figures are for all three fabs across two states, and construction jobs are by definition temporary.
There is also a fact that sits awkwardly beside the growth story and should not be omitted: TI conducted a round of layoffs at the existing Lehi fab, which the company characterized as organizational changes supporting its long-term operational plans. A site can be expanding and restructuring at the same time. Both are true, and a guide that reports only the first one is not a guide.
What actually happens inside, in plain terms
It helps to know roughly what the building does, because "chip factory" conceals a genuinely strange process.
A fab does not assemble chips. It prints them, in layers, onto a polished silicon disc. The disc is coated with a light-sensitive film, a pattern is projected onto it through a mask, the exposed film is developed away, and the material underneath is either etched out or has new material deposited into it. Then the disc is cleaned and the whole sequence runs again for the next layer. A finished part may involve dozens of these cycles, each one aligned to the last within a tolerance measured in nanometers.
Only at the end is the disc cut into individual rectangles — the die — which are tested, discarded if faulty, and sent onward to be packaged into the black plastic parts people recognize as chips. That packaging step frequently happens on a different continent from the printing step, which is one reason "made in America" is a harder claim to make about a semiconductor than about almost any other manufactured object.
Two consequences follow. A fab is fantastically sensitive to dust, vibration and contamination, because a single stray particle can ruin a die. And a fab has almost no useful intermediate output: it either yields good die or it does not.
What a fab needs from a valley
Fabs are not ordinary factories and their demands on a place are unusual.
They need enormous quantities of ultrapure water, which is why water recycling rates appear in the press releases. They need very stable power — a momentary sag can ruin work in progress across an entire tool. They need vibration isolation, which is why the buildings are massively constructed and sit on engineered slabs. And they need a supplier ecosystem within reasonable reach, which is the quiet argument for putting the second building next to the first rather than in a new state.
They also need people who can run them, which is the part that touches Utah County most directly.
The schools angle, and a timing wrinkle
TI committed $9 million to Alpine School District to expand science, technology, engineering and maths programs from kindergarten through twelfth grade. That is a long-horizon investment in exactly the workforce a fab needs and cannot import at scale.
There is a wrinkle worth flagging. Alpine School District is being replaced: on 1 July 2027 it splits into three districts, with Lehi falling into Aspen Peaks. How a multi-year corporate commitment made to a district that will no longer exist carries across to its successor is a reasonable question, and one nobody has publicly answered. See our Alpine School District guide for the mechanics of the split itself.
If you want to work there
Fab work is not a single job. The site needs process engineers and equipment technicians, but it also needs facilities, cleanroom operators, quality, planning, logistics and maintenance staff — and the technician layer is the largest and the most accessible.
Two honest observations. First, semiconductor manufacturing runs continuously, which means shift work, and shift work in a valley where the social calendar is heavily weekend- and evening-shaped is a genuine lifestyle decision rather than a detail. Second, the entry point most people miss is the technician route through a two-year program rather than the engineering route through a four-year degree.
Our guide to the biggest employers in Utah Valley sets the fab in context against the county's other large employers, and our jobs without a degree guide covers the manufacturing route in more detail.
Why it stays invisible
There is no sign you would notice, no visitor center, no tour and no retail frontage. Fabs are sealed by necessity: the cleanroom inside runs to airborne particle standards that make an operating theater look casual, and every visitor is both a contamination risk and a confidentiality risk.
That invisibility is why a valley that can tell you the founding date of every pioneer-era mill mostly cannot tell you what the biggest industrial building on its northern edge does. It is also why the Silicon Slopes story, as usually told, is a software story — Adobe's campus, the SaaS companies, the funding rounds — when the largest capital investment in the corridor by an order of magnitude is a manufacturing one.
The short version
A joint venture built a memory fab in Lehi in 2006. The product it eventually bet on did not sell. One partner bought the other out, then changed its mind within two years. A third company bought the building for $900 million, spent a year converting it, and began making an entirely different class of chip in December 2022. It then committed roughly $11 billion to a second building next door, drew up to $1.6 billion in milestone-based federal funding, and made Lehi one of the named sites in America's semiconductor policy.
The building did not change. What the country needed from it did.
Related Guides
- Silicon Slopes, explained
- Living in Lehi
- The biggest employers in Utah Valley
- Jobs in Provo without a degree
- Internships and careers in Provo
- Companies founded in Utah Valley
Frequently Asked Questions
Who owns the chip plant in Lehi, Utah?
What does the Lehi fab actually make?
What is LFAB1 and LFAB2?
How much is the Lehi expansion worth?
How many jobs will the new fab create?
What is a 300-millimeter wafer and why does it matter?
Are the chips made in Lehi advanced?
Can you tour the Lehi fab?
