Owners running hyperscale and colocation builds increasingly pre-purchase long-lead electrical and HVAC equipment (generators, switchgear, chillers, UPS modules) under OFCI (Owner-Furnished, Contractor-Installed) or OFE (Owner Furnished Equipment) terms, shifting procurement away from the general contractor and onto their own strategic sourcing desks [S2][S3].
Drivers are concrete: cloud-infrastructure spend passed $30 billion in a single quarter back in Q2 2020 [S2], the global data center generator market was projected to reach 10.9 billion by 2027 at roughly 6.5% CAGR [S4], and a single hyperscale site this year required offloading and setting more than 150 outdoor generators [S4]. Against that volume, the OFCI route is treated less as a contracting curiosity and more as a portfolio-level supply-chain program.
What OFCI and OFE actually mean on a data center project
OFCI is the acronym for Owner Furnished, Contractor Installed, and it covers any owner-supplied equipment or material that the contractor installs, from specialized process machinery down to structural steel and aggregate [S3]. On a data center, the OFCI bucket typically aggregates generators, switchgear, chillers, and other integrated power and cooling skids that have the longest manufacturing lead times [S3][S4].
OFE (Owner Furnished Equipment) is used loosely as a synonym in some procurement documents, but in practice it points at the same hand-off: the owner sources and pays for the asset, then hands it to a GC or MEP sub for receiving, storage, setting, and connection [S5][S4]. The split-responsibility point is exactly where most schedule risk lives, because the GC owns installation while the owner owns delivery, and any handoff gap lands on the critical path.
The classic OFCI workflow is: owner selects vendor, negotiates, places the order, tracks shipment, clears customs; contractor takes over at site with offloading, storage, inspection, and installation [S5]. Subcontracting portions of owner-side logistics to an owner's rep, a 3PL/4PL, or even back to the GC is common, but the contractual liability for timely delivery still sits with the purchasing party [S3].
OFCI versus CFCI: decision criteria for the procurement model
CFCI (Contractor-Furnished, Contractor-Installed) is the alternative where the contractor sources, pays for, and installs the equipment, giving the owner a single point of responsibility at the cost of visibility [S5]. The two models trade off across four well-defined axes: cost transparency, lead-time control, risk allocation, and administrative burden [S5].
The trade-off table is sharp enough to be useful on a one-page procurement memo. OFCI gives high cost transparency (direct vendor pricing, no GC markup) and lets the owner manage early orders against the critical path, but the owner carries high administrative burden and shares delivery risk with the installer. CFCI flips every cell: low cost transparency because contractor markups are bundled, but the contractor is centrally accountable and the owner's admin load collapses [S5].
For hyperscale owners with a strategic sourcing team and a repeatable basis of design, OFCI wins on standardization, pre-positioning, and bulk pricing. For a single-build owner with no internal procurement bench, CFCI wins on simplicity. The decision is rarely ideological, and industry guidance treats OFCI as a normal delivery model, with risk allocation shaped by the underlying contract form [S8].
Why pre-purchase beats the schedule: the lead-time math
Critical electrical and HVAC equipment for a 50-100 MW hall routinely runs long order-to-site lead times that exceed most greenfield build windows if the GC is placing the first PO at groundbreaking. OFCI lets the owner place that PO 6-12 months earlier, often before design is fully frozen, and park the equipment in a manufacturer-staging or owner-staging yard until the foundation is ready [S1][S2].
Strategic sourcing under OFCI converts a tactical, project-level procurement exercise into a portfolio-level partnership with the OEM, letting the owner aggregate demand across multiple sites, pre-position standardized skids regionally, and pull from inventory when a site hits its MEP readiness milestone [S2]. This is the same logic that makes modular skid fabrication viable: pre-fab the building subset plus the integrated IT, electrical, and mechanical gear in a factory, then ship a complete module to site for setting [S2].
The supply chain is not gentle. One hyperscale partner reported competing demand for thousands of generators globally, with production and trucking schedules slipping on the same call, sometimes hours before a planned arrival [S4]. Project buffering (deliberately holding vendor and crew capacity ahead of need) is one mitigation; another is locking OEM production slots via an OFCI master purchase agreement that pre-commits volume across the owner's pipeline.
Risks that come with the OFCI bargain
The first risk is liability: if the OFCI tile, transformer, or chiller shows up late, the owner pays for the schedule slip in the form of extended GC general conditions, idle heavy-lift crews, and stacking of trades on site [S3][S4]. Heavy-haul and rigging partners describe learning of delivery delays only hours before the planned arrival, which forces standby time and site rework as crews shuffle to knock out adjacent scope [S4].
The second risk is design freeze. OFCI requires the owner to commit to a specification, vendor, and configuration before the A&E team has finished the coordinated model, and late design changes (a bus rating bump, a chilled-water delta-T tweak, a switchgear UL 1066 vs IEC 62271 arc-flash re-spec) cascade back through purchase orders that are already in production [S1]. Pre-purchasing before design is finalized is a known hazard, and the standard guidance is to integrate delivery timelines directly into the design schedule rather than treating procurement as downstream of it [S1].
The third risk is interface ownership. Receiving, storage, preservation (desiccant, nitrogen-pad, or space-heater maintenance on transformers and switchgear), and setting all need a named responsible party in the contract; without that, the equipment can sit on a lowboy for weeks while two sides argue over who pays the demurrage. Industry guidance makes the point that risk allocation under OFCI is heavily shaped by the underlying contract, with turnkey contractors typically absorbing delivery risk in exchange for a markup [S8].
Contract structure: design-build, EPCM, and the OFCI carve-out
Under a design-build delivery, the owner keeps more design oversight and often retains the long-lead equipment buy directly, which is the natural habitat for an OFCI carve-out [S6]. Under EPCM, the engineering, procurement, and construction manager acts more as an agent, and OFCI patterns are equally common, with the EPCM providing procurement services on behalf of the owner rather than buying the asset themselves [S6].
For an owner considering a strategic-sourcing build-out, the practical recipe is: (1) standardize a basis of design for the top 10-20 OFCI line items, (2) negotiate master purchase agreements with 2-3 OEMs per line item with pre-committed volume, (3) tie the GC's contract to OFCI hand-off milestones with explicit receiving, storage, and preservation scopes, and (4) fund a small owner's rep or 3PL team to track shipments from factory FAT through site set [S2][S3]. Industry guidance frames OFCI as a normal, well-understood model rather than a niche contract structure [S8].
The workforce signal is also trackable: large owners are staffing dedicated strategic sourcing manager roles specifically for owner-furnished equipment on industrial compute campuses, which is a strong indicator that OFCI is moving from project-by-project firefighting to a permanent organizational function [S7].
When OFCI is the wrong tool
OFCI is the wrong procurement model when the owner lacks the internal bench to manage vendor selection, contracts, logistics, and customs, because the model concentrates rather than distributes risk [S5]. It is also the wrong model for small one-off builds, where the fixed cost of a sourcing team, a 3PL, and a customs broker exceeds any bulk-purchasing discount the owner could extract.
It is the wrong model for equipment with short manufacturing lead times (rack PDUs, standard breakers, commodity cable) where the GC's procurement leverage is already adequate. And it is the wrong model for any line item where the design is genuinely not stable, because the rework cost on a factory-staged asset can erase the schedule gain from pre-purchasing [S1][S3].
Trackable signals to watch over the next 12 months
Three signals are worth monitoring on a 2026-2027 horizon. First, hyperscale job postings for strategic sourcing managers with OFE or OFCI in the title, which indicate whether the model is becoming a permanent org layer rather than a project ad hoc arrangement [S7]. Second, the spread between factory-staging storage days and site-ready days on OFCI transformers and switchgear, which is the cleanest proxy for how well owners are syncing their OEM POs to the GC's critical path. Third, the appearance of standardized OFCI equipment specifications in industry reference documents, which current guidance already treats as a default rather than a deviation [S8].
Owners running a repeatable build program should expect OFCI to keep absorbing more of the long-lead scope, and contractors should expect to spend more of their project management hours on receiving, preservation, and hand-off coordination than on direct procurement, per the iRecruit procurement guide [S5]. For a deeper look at how pre-positioned equipment interacts with the building itself on a fast-track program, the modular construction and pre-fab data center workflow is a useful adjacent read, and the long-lead generator and switchgear market read provides the supply-side context. Engineers sizing the receiving and storage plan should also reference the heavy-lift and rigging math for pre-positioned skids when sequencing the OFCI hand-off.
For component-level specifications, see construction machinery and equipment, lighting equipment and electric lamps, and data logger.