Every major equipment category covered in this article, with current lead time ranges, demand pressure by item, RFS schedule risk, and the most common planning errors. 6-page PDF — save or share for baseline reviews and schedule update meetings.
Download PDF ReferenceThe Schedule That Looks Clean But Isn't
Most data center project delays don't start in the field. They start 18 months earlier — with a submittal log no one is watching closely enough.
I've reviewed schedules on hyperscale campuses where the critical path runs cleanly through structural steel, MEP rough-in, and commissioning. Logical sequence. No obvious red flags. But buried in the procurement section: medium voltage switchgear with a 52-week manufacturing lead time, submittal issued at month four of a 22-month program. The schedule assumed a 10-day AHJ review. The jurisdiction actually takes six weeks. Planned delivery lands at week 70. Actual delivery lands at week 74.
Four weeks of slip that no one tracked — and it surfaces at a schedule update review 14 months in, when there's no recovery path that doesn't cost someone a seven-figure acceleration premium.
And that's before you get to the transformer.
Power Is the Constraint
Nearly half of all respondents to the Turner & Townsend 2025–2026 Datacenter Construction Cost Index cited power access as the single biggest scheduling constraint — ahead of labor, land, and permits. Grid connection wait times in some U.S. markets now stretch to seven years. Transformer lead times have gone from 24–30 months pre-2022 to 80–128 weeks on average today, with large custom units reaching 3–5 years.
This is not a temporary disruption. Hyperscale AI buildout, grid modernization, EV infrastructure, and renewable integration are all competing for the same transformer manufacturing slots, the same grain-oriented electrical steel, and the same generator engine blocks. The constraint is structural, and it is not going to resolve within the delivery window of any program currently in planning.
On hyperscale programs, the energization date is fixed. The utility interconnect doesn't move. The IT operations team doesn't move their cutover. The only thing that moves is margin and reputation — and the schedule that was baselined with catalog lead times and no procurement risk narrative.
Equipment Risk Tiers
The equipment items with the highest schedule exposure on hyperscale data center programs fall into three risk tiers based on lead time, demand pressure, and impact on the RFS critical path.
Tier 1 — RFS-Critical
Large Power Transformers (50 MVA+): 80–128 week lead times, with large custom units reaching 3–5 years. Constrained by grain-oriented electrical steel (GOES) supply and a global manufacturing base concentrated in fewer than six producers. A typical 50 MW AI campus requires 4–6 of these. Procurement must open within the first 60 days of a program, not after design development.
Medium Voltage Switchgear (5–38 kV): 44–65 week lead times. Constrained by metal-clad production above 15 kV, SF6-free configurations, and AHJ review cycles that routinely add 4–12 weeks beyond the standard 10-day assumption. A 6-week delay here cascades 8–10 weeks into commissioning due to sequential energization requirements.
Diesel Generators (2 MW+ paralleled sets): 52–80 week lead times. The binding constraint is engine block allocation at Caterpillar, Cummins, and MTU/Rolls-Royce — not the packager's quote. These are frequently different numbers by 8–16 weeks, and the packager often doesn't know until the order is placed.
Tier 2 — High Risk
UPS Systems (large-frame modular): 40–72 week lead times. The frame is often not the constraint — power module allocation is. Projects that confirm UPS frame delivery may later learn modules are 12–16 weeks behind. Batteries are a third, independently procured line item that must be tracked separately.
Chillers (centrifugal/screw, 500+ ton): 40–60 week lead times. Constrained by compressor allocations and refrigerant transition complexity (R-410A phase-down, R-454B conversions). Specifying a refrigerant without confirming manufacturer production availability at the required tonnage is a current active risk.
Utility Substation Equipment: 40–78 weeks for the equipment alone, separate from the utility interconnect queue, which is not contractor-controlled. In some markets, the utility queue adds years beyond equipment lead time.
Bus Duct/Busway: 30–52 weeks. Often treated as a commodity item, but custom runs require final dimensioned drawings before fabrication can begin — a frequent source of procurement slippage.
Tier 3 — Elevated Risk
Cooling Distribution Units (CDU): 26–52 weeks, with demand surging 156% year-over-year in Q2 2025 per Dell'Oro Group. Historically treated as late-stage IT items, but on AI-density campuses CDUs are MEP critical path items that must be installed, piped, charged, and tested before GPU servers can be energized.
Dry-Type Distribution Transformers, ATS units: 20–40 week range. Risk compounds with quantity — one late unit can delay a full data hall.
Why AHJ Approval Is a Hidden Float Consumer
Equipment lead time is only half the exposure. On medium voltage switchgear, automatic transfer switches, and generator installations, the Authority Having Jurisdiction (AHJ) must review and approve submittals before manufacturing can begin or equipment can be commissioned. In many jurisdictions, AHJ review cycles run 30–45 days — not the 10-day assumption that appears in most construction schedules.
A submittal issued at month four with a 10-day AHJ assumption that actually takes 42 days adds four weeks to the procurement start before the manufacturer's lead time clock even begins running. On a 52-week lead-time item, that converts a planned week-70 delivery into a week-74 delivery — with no float to absorb it if the item is on the critical path.
This is a recoverable risk if it is identified at baseline. It is not recoverable if it is discovered at month 14.
What a Credible Schedule Shows
On a hyperscale program, a credible baseline schedule includes:
- Vendor-confirmed manufacturing slots — not catalog lead times — for every Tier 1 and Tier 2 item
- AHJ submittal-to-approval durations based on jurisdiction history, not standard assumption
- UPS frame, modules, and batteries tracked as separate procurement line items with independent delivery dates
- Generator engine block allocation confirmed separately from the packager's quote
- CDUs classified as mechanical (MEP) items with CPM logic-links to the chilled water commissioning sequence
- Utility interconnect duration treated as a variable, not a fixed date
If any of these elements are missing from the current schedule, the float on the critical path is not real — it is an artifact of procurement assumptions that haven't been tested against the supply chain.
The downloadable reference paper above covers each equipment category in full, with common planning errors documented by item and a three-phase procurement risk checklist (baseline, monthly update, 90% design). CPM Pros runs procurement risk reviews on every hyperscale engagement — if your program is entering baseline development or in schedule recovery, independent schedule peer review delivers a written report within 10 business days.