If you're asking who is building new chip factories in the USA, you're not just looking for a list of names. You want to know the scale, the locations, the technologies, and—let's be honest—whether this massive wave of investment is more than just headlines. Having followed semiconductor construction projects from blueprint to groundbreaking, I can tell you the reality on the ground is more complex and more fascinating than the press releases suggest. It’s not just about Intel and TSMC. A whole ecosystem is being rewired, from Arizona's desert to Ohio's farmlands, and the map is being redrawn in real time.
What’s Inside This Guide
The Major Players: A Complete Factory List
Forget the vague summaries. Here’s the concrete, project-by-project breakdown of who is building what and where. This table consolidates information from company announcements, state filings, and industry reports like those from the Semiconductor Industry Association.
| Company | Project Name / Location | Investment (Estimated) | Technology / Focus | Status (as of latest updates) |
|---|---|---|---|---|
| Intel | Ohio One (New Albany, OH) Fab 52 & 62 (Ocotillo, AZ) Expansion in Rio Rancho, NM |
$20B+ (OH initial phase) $20B (AZ expansion) $3.5B (NM) |
Advanced logic (Intel 18A/20A), leading-edge foundry services | Construction underway in OH & AZ. NM expansion in planning. |
| Taiwan Semiconductor Manufacturing Company (TSMC) | Fab 21, Phase 1-3 (Phoenix, AZ) | $40B+ (total planned) | Advanced logic (4nm, 3nm, potentially 2nm) | Phase 1 (N4) equipment move-in. Phase 2 (N3/N2) under construction. |
| Samsung Electronics | New Fab (Taylor, TX) | $17B+ | Advanced logic for mobile, HPC, and foundry customers | Steel structure complete, targeting 2024 production. |
| Micron Technology | Leading-Edge Memory Megafab (Clay, NY) Expansion in Boise, ID |
$100B+ over 20+ years (NY) $15B (ID) |
DRAM memory chips (1-gamma node and beyond) | Site preparation in NY. ID fab construction planned. |
| Texas Instruments | RFAB2 & LFAB (Sherman, TX) Expansion in Lehi, UT |
$30B+ (TX sites) $11B (UT) |
Analog and embedded processing chips (300mm wafers) | Multiple fabs under construction in Sherman. |
| Wolfspeed | John Palmour Manufacturing Center (Siler City, NC) | $5B | Silicon Carbide (SiC) wafers for EVs and energy | Construction underway, first tools installed. |
| GlobalFoundries | Expansion at Malta, NY campus | $1B (initial phase) | Feature-rich and specialty technologies (RF, automotive) | New building constructed, equipment installation phase. |
Driving through the industrial parks near Phoenix or Austin, the scale hits you. It’s not just a building; it’s a small city of cranes, concrete, and specialized contractors. The chatter at local supplier meetups confirms this: the activity is real, but the timelines are fluid. A common misconception is that these are "copy-paste" fabs from Asia. They're not. The designs are often adapted for higher automation, different utility specs, and future expansion, which adds complexity.
What's Really Driving the Building Boom?
Everyone points to the CHIPS and Science Act. That's the catalyst, but it's not the whole story. The funding is a powerful incentive, but these companies aren't spending tens of billions just for a government grant. The strategic calculation runs deeper.
The Geopolitical Imperative
The pandemic chip shortage exposed a terrifying vulnerability. A single geopolitical event or natural disaster in East Asia could halt global production of everything from cars to medical devices. Building in the USA is a massive insurance policy. For TSMC and Samsung, it’s also about pleasing their largest customers—Apple, AMD, Nvidia, Qualcomm—who are desperate for geographic diversification in their supply chains. I've heard from engineers at these fabless companies that the pressure to have a "non-Taiwan" option for advanced nodes is now a top-tier boardroom discussion.
The Economic and Technical Calculus
While labor and upfront costs are higher in the U.S., the total cost of ownership picture is changing. Proximity to R&D centers (like Intel’s in Oregon or tech hubs in California) can accelerate innovation cycles. For chips going into defense, aerospace, or critical infrastructure, a "Made in USA" tag is becoming a prerequisite, not a luxury. The investment in materials companies like Wolfspeed shows the push isn't just about leading-edge logic; it's about securing the entire stack for next-gen applications like electric vehicles.
Why These Specific Locations Were Chosen
The map isn't random. It’s the result of a brutal site selection process weighing dozens of factors. Water, power, talent, and logistics are the big four.
Arizona (Intel, TSMC): It’s not just about sunny weather. The state has a deep history in semiconductor manufacturing, meaning a pre-existing ecosystem of suppliers, a trained technical college network, and governments that understand the utility needs (immense amounts of ultra-pure water and stable, gigawatt-scale power). The water issue is critical and often underplayed—these fabs have multi-million-dollar water reclamation facilities built in.
Ohio (Intel), New York (Micron, GlobalFoundries), Texas (Samsung, TI): These are plays for long-term talent and space. These states offered massive tracts of contiguous land (you need room for 4-5 fabs over decades), significant utility infrastructure commitments, and strong partnerships with universities to pipeline engineers. Ohio’s pitch, for instance, wasn't just about tax breaks; it was about co-investing in curriculum at Ohio State and community colleges to create a homegrown workforce.
The Unspoken Challenges and Real Impact
Construction is the easy part. The hard part comes next.
The Talent Gap: We simply don't have enough skilled technicians, process engineers, and fab managers in the U.S. to staff all these new factories overnight. TSMC has faced cultural and logistical hurdles bringing experienced engineers from Taiwan to Arizona. The solution is a massive, nationwide effort in technical education, which has started but will take years to bear fruit.
The Supplier Gap: A fab isn't an island. It needs a network of hundreds of nearby suppliers for specialty gases, chemicals, wafer handling equipment, and maintenance. Much of that ecosystem is still in Asia. Building it stateside is a parallel, less glamorous investment that’s just as crucial.
The Realistic Timeline: Don't expect these fabs to solve near-term chip shortages. From groundbreaking to high-volume production of advanced chips can take 3-5 years. The initial impact is in construction jobs and local economic boosts. The strategic supply chain resilience benefits will materialize in the latter half of this decade.
The impact, however, is profound. It’s reversing a decades-long trend of offshoring. It’s creating hubs of advanced manufacturing that will spawn innovation and secondary businesses. For a town like Taylor, Texas or Clay, New York, it’s a generational transformation.
