The Equation Series
Menu

Industry study 07 of 24 · Manufacturing Research Companion: Companies, Strategic Moves, and Technical Studies

Plants, Site Selection, Capex and Ramp-Up

~7,000 words

A plant is the largest, longest and least reversible decision a manufacturer makes, and the gap between announcing one and running it at volume is where much of the value in this book is created or destroyed. The fastest modern example is Tesla Manufacturing's Shanghai plant: groundbreaking on 7 Jan 2019 and first customer deliveries of the Model 3 on 30 Dec 2019 (equal to Energy: Tesla's Giga Shanghai move). The slowest are chip fabs: Intel Foundry broke ground in Ohio in September 2022 (from memory, verify date) and has since moved the first fab's completion to 2030 with operations in 2030–31 (company, Feb 2025, reported); TSMC Arizona's first fab reached volume production on N4 in late 2024 (reported), shipped its first Blackwell wafer in Oct 2025, sits inside a $165bn US plan (Mar 2025), and its second fab is now expected to reach 3 nm volume in the second half of 2027 (TrendForce, reporting TSMC's Apr 2026 call), a year or more later than the 2026 originally announced; Samsung's Taylor, Texas fab, announced in 2021 and backed by a $4.745bn CHIPS award, obtained temporary certificates of occupancy in 2026 and was preparing to make chips under a reported ~$16.5bn Tesla contract, with the company saying "start of production" meant readiness to operate by end-2026 (reported). Between the fast and the slow are the 2022–26 battery plants, many cut back or converted, and the closures — GM's Lordstown in 2019, sold to Lordstown Motors, then to Foxconn (~$230m, closed May 2022), then to a SoftBank entity (~$375m including equipment, Aug 2025; equal to Supply chain: Foxconn). This section explains greenfield, brownfield and expansion choices, how sites are chosen, what plants cost, why megaprojects slip, how ramps and yield learning work, the construction-labour constraint, the difference between "announced" and "operating", and what closures cost the towns around them. It has no worked example: the main book's Appendix A computes one new US plant's NPV with a grant, a clawback and a production credit, and this section links to it rather than repeating it.

What the study covers

  1. 1Overview: the decision that outlives its makers
  2. 2Units
  3. 3Greenfield, brownfield, expansion and conversion
  4. 4Site selection
  5. 5What plants cost
  6. 6The plant timeline: from announcement to volume
  7. 7Megaprojects and their delays
  8. 8Ramp-up curves and yield learning
  9. 9Construction labour and the build constraint
  10. 10Announced vs operating
  11. 11Closures and community costs
  12. 12Europe and Asia
  13. 13How to read a plant announcement
  14. 14Key numbers
  15. 15Landmark companies and cases
  16. 16Pitfalls
  17. 17How an expert thinks about it
  18. 18Glossary

Companies in this study