Analytics24
3 September
My list

The Long Read · 22 Aug

South Boston ships in 2029

South Boston ships in 2029. The yard date is the product; the slide that says sooner is not.

A crane hook and a folded height variance. A bay that has not shipped.

Every leftover from the last sink sentence has a different object. Hitachi Energy named a $457 million expansion in South Boston on 29 June. Hefei is a 17 August bushing shop. Butler Works is one mill in Lyndora. Flatten those into one domestic transformer trade and you get a false headline. Do not flatten them.

2029

Hitachi Energy groundbreaking, South Boston, 29 Jun 2026. $457m, ~825 jobs, campus since 1968. ENR: >300k sq ft, ~130 ft, 375-ton crane. Callahan: a little behind. Target completion 2029.

A 2027 hall is waiting on a crane that has not been hung.

Groundbreaking, crane, railcar

29 Jun 2026Groundbreaking$457m · 825 jobsCrane bay375-ton hookNot hung2029 railcarNameplateNot a 2027 hall
The press release does not ship a unit. Existing bays already do. The expansion answers 2029, not a 2027 service date.

The dates are already on paper. 29 June on a Halifax groundbreaking. 2029 on a Virginia completion. 17 August on a Hefei note. 2028 on a Butler induction furnace.

South Boston is the exposed bay

The letter the catalog actually produced is a crane hall, not a 2027 interconnect. Hitachi Energy, 29 June 2026, South Boston, Virginia: the nation's largest facility for large power transformers. $457 million. About 825 jobs on about 850 already there. The campus has been a plant since 1968, when it was Westinghouse. Governor Spanberger, Senator Warner, Congressman McGuire. The press release does not ship a unit.

Engineering News-Record, on the same project: a manufacturing building of more than 300,000 square feet, assembly bays taller than 120 feet, overall height about 130 feet, a 375-ton overhead crane, rail out the back for oversized loads. Halifax County had to grant a height variance because the old zoning was not written for a hall that tall. Greg Callahan, who runs the North American transformer book, told ENR the project was "a little behind where we want it to be." Completion is still targeted for 2029. A 2027 AI hall that has not ordered into an existing bay is waiting on a crane that has not been hung.

CISA, under Homeland Security, has already said the shortage of large power transformers is a grid-reliability problem, not a procurement inconvenience. Lead times that sat near 50 weeks in 2021 were printed around 120 weeks by 2024 in that material. Industry trackers in 2026 talk about 128 weeks and, for the biggest units, three to five years. A GPU can be allocated this quarter. A large-power transformer is a serial object: electrical steel, a winding, a dry-out, a test bay, a railcar. You cannot hang a second crane inside the same bay and call it parallelism.

The power page already named a PJM tariff, an ERCOT request pile, a GE Vernova slot into 2031. Those are real. They are also upstream of a winding. Hitachi's own sentence is larger than Virginia. The South Boston slab is a cornerstone of more than $1 billion of US grid-equipment manufacturing. On 17 August the same company put $300 million into Hefei for transformers, ultra-high-voltage bushings, and tap changers, and called that part of a $9 billion global plan. It is one bay in a book that still has to import a bushing. Do not read a groundbreaking as domestic self-sufficiency. Read the first unit that leaves on the rail with a nameplate and a customer.

Existing bays already ship. This piece does not say no transformer will leave America before 2029. It says the expansion that was announced as the answer to the AI load does not answer it on the 2027 service dates the campuses still print.

What a groundbreaking cannot do

Cross-asset does not care which crane is hung. AI abundance cares only if the middle's hosting cost becomes a steel-and-crane cost. It has not, on the invoices those pages already named. The sink theme already named Hefei, Butler, Foundayo, and EMBRAZE. The bushing piece already named the lid. This piece is the bay.

A refutation would look like this: a named large-power transformer from the South Boston expansion, with a Hitachi or customer notice, by 31 December 2029, confirms the slab shipped a unit. A 2029 that is still a height variance and a hiring notice confirms the 2027 hall bought a slot in a plant that was not yet a plant. Eaton and GE Vernova are the listed peers the piece is allowed to point at. They are not this crane.

The honest uncertainty is which object gets a number first. If a named South Boston expansion unit leaves on the rail and Hefei is still a spend, Virginia moved and the insulator did not. If Hefei ships a named bushing and South Boston is still a crane, the insulator moved and Virginia did not. If groundbreakings keep coming and no railcar appears, the $457 million stayed a tour. Those three sentences can live in the same year. Four hundred and fifty-seven million dollars are a fact about one bay. Three hundred million dollars are a fact about one shop. Neither is a winding on a railcar.