Utility infrastructure decisions made in basic engineering phase determine 5–10x cost ratios between optimal and retrofitted outcomes. AI-ready manufacturing facilities now require 2–3x higher electrical capacity than traditional plants, water systems must support 24-hour continuous production, and time-to-power has become the single most decisive constraint in greenfield site selection. This checklist covers seven utility systems with 52 specific items to lock before breaking ground — structured exactly the way capital project teams approve and execute them. Schedule a utility planning consultation to walk through this checklist against your specific facility design and timeline.
Greenfield Utility Infrastructure Checklist · 2026
7 Utility Systems · 52 Checklist Items
Power, water, natural gas, telecom, compressed air, steam, and wastewater — the utility decisions that determine whether a greenfield plant comes online on schedule or slips into commissioning purgatory.
01
Electrical Power
12–138 kV substation
10 items
02
Water Supply
Potable + process + fire
8 items
03
Natural Gas
2–60 PSIG supply
6 items
04
Telecom & Network
Fiber + 5G + redundant
8 items
05
Compressed Air
ISO 8573-1 quality
7 items
06
Steam Generation
15–600 PSIG boilers
6 items
07
Wastewater
Pre-treat + discharge
7 items
2–3xelectrical capacity AI plants need
30%headroom above projected peak demand
5–15%rework cost when utility decisions slip
Why Utility Master Planning Matters in 2026
Three industry shifts have made utility infrastructure planning more decisive than at any point in the past two decades. AI-ready manufacturing facilities require electrical capacity that legacy planning frameworks dramatically underestimate. Regional utility availability has become the binding constraint in site selection. And reshoring is driving greenfield announcements at a pace that strains permit processing, construction labor, and equipment lead times across every utility category simultaneously. Plants that treat utility planning as a late-stage MEP decision rather than a basic engineering priority are locked into cost ratios they cannot recover.
01
AI Capacity Demand Surge
AI-ready manufacturing facilities require 2–3x higher electrical capacity than traditional plants. Edge computing, AI inference hardware, and robotics all carry significant power demand. Traditional greenfield electrical planning consistently under-provisions by 40–60% for AI-native operations.
02
Time-to-Power as Site Selection
In 2026, power availability and infrastructure delivery timelines are the most decisive factors in site selection. Some U.S. utility territories now quote 24–36 month timelines for new industrial transformer connections. Site selection without verified power availability creates compounding schedule risk.
03
Regulatory Complexity Climbing
Wastewater discharge permits, air quality permits, water rights, and natural gas supply agreements each have multi-year approval timelines. Coordinated permitting across utility categories is now a dedicated workstream in serious greenfield projects, not a back-office task.
04
Redundancy Required, Not Optional
Production downtime cost reaches $260K per hour for high-value manufacturing. Single point of failure on any major utility creates unacceptable risk. Greenfield plants now design dual-feed power, redundant water supply, and backup telecom from day one rather than retrofitting after first outage.
The 7 Utility Systems Checklist
The checklist below organizes 52 utility infrastructure items across seven systems. Best practice: complete this checklist during basic engineering phase, 12–18 months before mechanical completion target. Each item should have a documented owner, decision status, and dependency chain mapped before construction starts. Click each utility category below to expand the detailed checklist — items marked P1 are critical-path decisions that gate downstream commitments.
Walking through this checklist for your specific facility? Book a utility planning consultation — we’ll evaluate each utility category against your production requirements, geography, and timeline.
01
Electrical Power Supply
12–138 kV substation · Dual-feed redundancy
10 items
1.1
Peak demand calculation with 30% AI/automation headroom · P1
Calculate peak kW load including production equipment, HVAC, AI compute, and EV charging with 30% headroom. AI-native plants require 2–3x traditional capacity — legacy frameworks under-provision by 40–60%.
1.2
Substation primary voltage and capacity confirmed · P1
Industrial substations operate at 12kV, 25kV, 35kV, 69kV, or 138kV primary. Confirm capacity supports peak demand plus growth. Transformer lead times: 12–24 months.
1.3
Dual-feed or redundant utility connection · P1
Require dual substations or independent feeders. Single point of failure shuts down the plant. Mission-critical: two physically separate connections from different substations.
1.4
Power quality specification documented
AI inference hardware, VFDs, and robotics require clean power. Specify voltage stability (±5%), THD <5%, frequency stability. Evaluate local grid uptime and fluctuation history.
1.5
Time-to-power timeline confirmed by utility
Some U.S. utility territories quote 24–36 months for new industrial transformer connections in 2026. Get utility commitment in writing before site selection finalizes.
1.6
Backup generation sized (UPS + generator)
Edge computing, OT controls, and critical loads need uninterruptible power. UPS for 15 min full-load runtime, generator for extended outages. ATS for seamless cutover.
1.7
Renewable energy integration assessed
Evaluate rooftop solar, on-site wind, local green energy programs. IRA Section 48 ITC (30–50% with stacking) makes renewable integration a competitive advantage.
1.8
Distribution system & switchgear specified
Primary distribution (480V or 4160V), secondary (208V/120V), MCCs, and panel boards designed for production load. Switchgear supports maintenance without full shutdown.
1.9
Grounding & lightning protection designed
Grounding grid resistance specifications, lightning protection per NFPA 780, surge protective devices at key locations. Critical for sensitive electronics and insurance compliance.
1.10
Energy storage (BESS) integration planned
BESS for demand charge reduction (30–50% of bill), TOU arbitrage, demand response. Sized at 25–40% of peak demand. Design in basic engineering — retrofitting costs 5–10x.
02
Water Supply Systems
Potable + process + fire suppression
8 items
2.1
Water demand calculation by use category · P1
Calculate demand by category: process water, cooling tower makeup, boiler feedwater, CIP/SIP, potable, fire suppression, irrigation. Different quality and pressure per category.
2.2
Water supply source confirmed (municipal vs well) · P1
Municipal connection capacity confirmed in writing, or groundwater well permits with verified yield and quality. Surface water requires state water rights. Multi-year permit timelines.
2.3
Water treatment requirements specified
Process water treatment for hardness, decarbonation, dechlorination, RO/DI as required. Boiler feedwater per ASME. Pharma/F&B may need ultrapure (Type 1, 2, or 3).
2.4
Storage tank capacity sized
On-site storage sized for 1–3 days continuous operation. Separate tanks for potable, process, and fire suppression. Fire storage per NFPA 22: 250,000–500,000 gallons typical.
2.5
Fire suppression water system designed
Fire pump capacity per NFPA 20, hydrant spacing per fire marshal, sprinkler density per occupancy. Insurance carrier (FM Global, HSB) may exceed code minimums.
2.6
Distribution piping & pressure zones specified
Pipe sizing for flow + pressure loss, materials per chemistry (CS, SS, PVC, HDPE), pressure zone segregation, isolation valves at branches. 40–80 PSI at point of use.
2.7
Backflow prevention & cross-connection control
RPZ backflow preventers between potable and process, double-check valve assemblies at irrigation, air gaps where required. Annual testing. State health approval typically required.
2.8
Water reuse & sustainability designed
Cooling tower blowdown recovery, CIP rinse reuse, condensate return, rainwater harvesting. Reuse reduces makeup demand 20–40%. ESG reporting requires water intensity metrics.
03
Natural Gas Supply
2–60 PSIG · Metering · Safety systems
6 items
3.1
Natural gas demand & pressure required · P1
Calculate peak SCFH from process equipment, boilers, furnaces, ovens, makeup air. Pressure: 2–15 PSIG for boilers, up to 60 PSIG for industrial process.
3.2
Gas main capacity & availability confirmed · P1
Utility commitment for required volume in writing. Constrained service territories require expensive upgrades or alternative fuels. Confirm before site selection finalizes.
3.3
Regulator station & metering designed
Pressure regulation from main (60–200 PSIG) to plant distribution. Metering for billing and consumption tracking. Over-pressure protection per ANSI B31.8.
3.4
Distribution piping designed & sized
Pipe sizing for flow + pressure drop, welded steel typical for industrial, isolation valves, leak detection. NFPA 54 and ASME B31.8 compliance. Coordinate structural pipe support and seismic.
3.5
Emergency shutoff & gas detection systems
Main building shutoff, emergency manual shutoffs at key locations, gas detection in confined spaces. BMS-integrated for automatic shutoff. Required by NFPA and local fire code.
3.6
Alternative fuel backup evaluated
For critical operations: propane backup, dual-fuel burners (gas + oil), or electrification. Gas curtailment events interrupt supply in some regions — critical loads need fuel diversity.
04
Telecom & Network Infrastructure
Fiber + 5G + redundant carriers
8 items
4.1
Fiber optic service from multiple carriers · P1
Diverse fiber paths from two+ carriers, physically separated building entry. AI-ready plants need 1–10 Gbps initial, scalable to 100 Gbps. Avoid single-carrier dependency.
4.2
Telecom room (MDF) design specified · P1
MDF room with dedicated HVAC (65–75°F), redundant power, fire suppression, security access. Cable entry from two+ directions for path diversity.
4.3
5G/4G LTE backup connectivity
Cellular backup for fiber outages via SD-WAN. Private 5G for high-density IoT and mobile equipment. Verify cellular coverage strength on site before construction.
4.4
Structured cabling plan documented
Cat6A/Cat8 for office and floor distribution, OM4/OS2 fiber backbone. IDF rooms at distribution points. Cable trays sized for current + 50% expansion. TIA-568, BICSI compliance.
4.5
Network architecture (LAN/WAN) designed
Core/distribution/access switches, OT/IT segmentation per IEC 62443, VLAN strategy, firewalls, wireless AP density. Industrial Ethernet (OPC-UA, PROFINET, EtherNet/IP).
4.6
Cybersecurity architecture per IEC 62443
Defense-in-depth: perimeter firewall, segmentation, OT/IT DMZ, intrusion detection, SIEM. Required by insurance and regulation. Designed in basic engineering, not bolted on.
4.7
Wireless coverage planning
WiFi 6/6E or 7 AP density for warehouse, floor, office. Bluetooth/UWB for asset tracking. RFID at receiving/shipping. Site survey before AP placement finalizes.
4.8
Edge data center / server room
On-premise compute for AI inference, MES, SCADA, video analytics. UPS power, redundant cooling, clean agent fire suppression, physical security. Typical: 200–1000 sq ft.
05
Compressed Air Systems
ISO 8573-1 quality classes · CDA
7 items
5.1
Air demand calculation by quality class · P1
Calculate SCFM by class: plant air (general pneumatic), instrument air (controls), Clean Dry Air (precision). Each class has separate generation and distribution.
5.2
Compressor type & redundancy specified
Centrifugal (continuous), rotary screw (most common industrial), reciprocating (intermittent). N+1 redundancy standard. VFDs improve part-load efficiency.
5.3
Air quality specification per ISO 8573-1
Class for particles, water (pressure dew point), oil. Plant air: Class 3-4-4. Instrument air: Class 2-3-2. CDA: Class 1-2-1.
5.4
Air dryers & filtration designed
Refrigerated dryers (plant air), desiccant (instrument air, low dew point), membrane (point-of-use CDA). Pre/after-filtration for water, particulates, oil. Bypass and isolation for maintenance.
5.5
Storage receivers sized
Wet and dry receivers dampen demand fluctuations and provide outage ride-through. Sizing: 1 gallon per CFM. Larger receivers reduce cycling and extend equipment life.
5.6
Distribution piping & pressure drop
Pipe sizing maintains pressure drop <10% from compressor to point of use. Galvanized steel, aluminum, or stainless. Sloped for condensate. Loop config for redundancy.
5.7
Compressed air leak management plan
Compressed air: $0.20–$0.40 per 1,000 SCF. Leaks waste 20–30% in untracked plants. Specify ultrasonic detection, leak tagging, remediation tracking from day one.
06
Steam Generation
15–600 PSIG · ASME-stamped boilers
6 items
6.1
Steam demand & pressure specified · P1
Calculate lb/hr by process equipment, CIP/SIP, HVAC, humidification. Pressure tiers: 15 PSIG (HVAC), 150 PSIG (process), 600+ PSIG (power gen). Demand profile drives boiler selection.
6.2
Boiler type, capacity, redundancy
Firetube (small/medium, lower pressure), watertube (high pressure, larger), electric (zero emissions). N+1 redundancy. ASME stamped. Typical greenfield: 2–4 boilers.
6.3
Feedwater treatment system
Softening, dealkalization, RO/DI per boiler pressure rating. Chemical treatment (oxygen scavengers, pH, scale inhibitors). Treatment failure causes tube failures — expensive incident category.
6.4
Condensate return system
Condensate collection and return to deaerator captures water and heat. Reduces makeup water 60–80%, fuel 10–20%. Specify in basic engineering.
6.5
Steam distribution & pressure reduction
Insulated piping per ASME B31.1, PRV stations for multi-pressure, steam traps, isolation/drain valves. Distribution losses 5–10% — properly insulated runs to point of use.
6.6
Boiler emissions & permitting
Air permits for NOx, CO, particulates. Low-NOx burners required in many jurisdictions. CEMS stack monitoring above threshold. Permit timelines: 12–24 months. Electrification reduces permitting burden.
07
Wastewater Treatment
Pre-treat + discharge permit
7 items
7.1
Wastewater volume & characterization · P1
Calculate daily and peak flow, characterize contaminants: BOD, COD, TSS, pH, oils/grease, heavy metals, industrial pollutants. Drives treatment design and discharge permit specs.
7.2
Discharge permit secured · P1
Industrial pretreatment permit (POTW), NPDES permit (surface water), or zero liquid discharge. Permit timelines: 6–24 months. Permit conditions drive treatment design.
7.3
Collection system designed
Gravity sewers and pressure forcemains sized for flow, manholes at direction changes, lift stations where required. Segregate process wastewater from sanitary/stormwater. Sample manholes for monitoring.
7.4
Pre-treatment system specified
Screening, equalization, neutralization, oil/water separation, DAF, biological treatment as required by permit limits. VFDs on pumps improve efficiency 20–40%.
7.5
Aeration system designed (biological treatment)
Fine bubble diffused aeration (1.8–2.5 kg O₂/kWh, highest efficiency) vs surface aerators (lower capital, lower efficiency). Aeration is dominant operating cost. Diffuser replacement: 8–12 years.
7.6
Sludge handling & disposal plan
Sludge thickening, dewatering (belt press, centrifuge), storage, disposal contracts. Disposal: $50–$300 per wet ton by classification/geography. Plan in basic engineering.
7.7
Monitoring & reporting systems
Continuous monitoring at discharge point per permit (pH, flow, conductivity, contaminants). Sampling/lab schedule. Monthly/quarterly regulatory reporting. Automation reduces compliance burden.
Working through this checklist for your specific facility? Schedule a utility planning session — we’ll walk through each category against your production requirements, geography, and timeline pressure.
Lock Utility Decisions Before Breaking Ground
A utility planning consultation evaluates each category against your specific production requirements, geographic constraints, regulatory environment, and ramp-up timeline. Output: a documented utility plan with capacities, redundancy strategy, and permit timeline.
Utility Redundancy Planning Framework
Single point of failure on any major utility creates unacceptable production risk for high-value manufacturing. Production downtime costs reach $260K per hour, which justifies higher redundancy than initial budgets often allow. The four redundancy patterns below cover the dominant approaches in 2026 greenfield facilities. Each utility category gets a documented redundancy strategy during basic engineering — retrofitting redundancy after first outage costs 5–10x.
Strategy N+1
One Spare
Single redundant unit covers failure of any one operating unit. Most common pattern for compressors, boilers, chillers, pumps. Cost-effective for utilities with multiple parallel units.
Best for: Compressed air, steam, water pumps, chillers
Strategy 2N
Full Duplication
Complete second system independently capable of handling full load. Used for mission-critical utilities where any interruption is unacceptable. Higher capital cost but lowest failure risk.
Best for: Electrical power, telecom, fire protection
Strategy 2N+1
Duplication + Spare
Two independent systems plus a third spare. Highest reliability tier — used for tier-4 data centers, semiconductor fabs, and continuous-process operations where any downtime is catastrophic.
Best for: Mission-critical AI compute, pharma cleanrooms
Strategy Diverse
Different Sources
Two utilities from physically separated and operationally independent sources. Dual utility feeds from different substations, two carriers for telecom, alternative fuels for combustion. Protects against utility-side failures.
Best for: Electrical primary feed, fiber, fuel supply
Designing redundancy strategy for your facility? Book a redundancy planning session — we’ll map appropriate redundancy patterns to each utility against your production downtime tolerance and capital budget.
Common Greenfield Utility Mistakes
Four utility planning mistakes account for the majority of avoidable greenfield delays and cost overruns. Each is recognizable in retrospect but easy to commit during basic engineering pressure. Recognizing them before commitment is the discipline that distinguishes successful greenfield projects.
01
Under-Provisioning Electrical Capacity
Traditional planning frameworks under-provision electrical capacity by 40–60% for AI-native operations. Edge computing, AI inference hardware, robotics, and EV charging all carry significant power demand. Once electrical service is sized, expanding requires utility upgrades with 12–24 month timelines.
02
Sequential Permitting
Treating permits sequentially (electrical first, then water, then wastewater, then air) extends critical path by 12–18 months unnecessarily. Modern greenfield projects coordinate permits across utility categories on an integrated timeline. Requires upfront coordination but compresses overall schedule meaningfully.
03
Underestimating Time-to-Power
2026 U.S. utility transformer connection timelines have stretched to 24–36 months in some territories. Site selection without verified power timeline in writing creates compounding schedule risk. Time-to-power is now the single most decisive site selection constraint.
04
Retrofitting Redundancy After Outage
Designing single-feed utilities to minimize initial cost, then retrofitting redundancy after first significant outage. Retrofit cost: 5–10x the cost of designing redundancy from day one. Production downtime cost of $260K per hour typically justifies redundancy upfront.
Expert Perspective
"The utility decisions made in basic engineering phase determine 5–10x cost ratios between optimal and retrofitted outcomes — and almost every greenfield project we evaluate has at least one utility category where these decisions need to be revisited. Time-to-power has become the single most decisive site selection constraint in 2026. AI-ready plants require 2–3x traditional electrical capacity that legacy planning frameworks dramatically underestimate. Wastewater permits take 12–24 months. Compressed air leaks waste 20–30% of generation in untracked plants. None of these are technology problems — they’re planning problems that compound when teams treat utility infrastructure as a late-stage MEP decision rather than a basic engineering priority. The plants that get utility planning right design redundancy from day one, coordinate permits across categories in parallel, size for AI-native demand profiles, and commit to redundancy strategies that match production downtime tolerance to capital budget. The plants that get it wrong are the ones building the case studies on what not to do."
— Greenfield Utility Practice, 2026 perspective
2–3x
AI-ready electrical demand vs traditional
24–36 mo
U.S. utility transformer timeline 2026
5–10x
retrofit cost vs basic engineering design
Build Your Greenfield Utility Master Plan
A utility planning consultation evaluates each of the seven utility systems against your specific production requirements, geographic constraints, and ramp-up timeline. Output: a documented utility master plan with capacities, redundancy, and integrated permit timeline.
Frequently Asked Questions
When should we start utility infrastructure planning?
Begin utility planning during basic engineering phase, typically 18–24 months before mechanical completion target. Electrical service availability and time-to-power should be verified before site selection finalizes (some U.S. utility territories now quote 24–36 month transformer timelines in 2026). Wastewater discharge permits should be initiated 12–18 months before construction kick-off. Sequential planning extends critical path; integrated planning compresses it. The basic engineering phase is the right window because facility design constraints, vendor selection, and procurement specifications can all incorporate utility decisions before they’re locked.
How much electrical capacity do AI-ready facilities really need?
AI-ready manufacturing facilities require 2–3x higher electrical capacity than traditional plants. Drivers: edge computing for AI inference (50–200 kW), AI vision systems on production lines, increased robotics density, EV charging for fleet vehicles, and higher HVAC load supporting compute infrastructure. Traditional greenfield planning frameworks under-provision by 40–60% for AI-native operations. The conservative approach: calculate steady-state and peak load, add 30% headroom above projected peak demand for growth and AI/automation. Under-provisioning costs more to fix than to prevent — expanding electrical service typically requires utility upgrades with 12–24 month timelines.
What redundancy level should we design for each utility?
Redundancy strategy is driven by production downtime cost and operational criticality. Standard tiers: N+1 (one spare) for utilities with multiple parallel units like compressors, pumps, and boilers; 2N (full duplication) for mission-critical utilities like electrical primary feed, telecom, and fire protection; 2N+1 (duplication + spare) for tier-4 critical operations like semiconductor fabs and pharma cleanrooms; Diverse Source (utilities from physically separated sources) for protecting against utility-side failures. Production downtime of $260K per hour typically justifies higher redundancy than initial budgets allow. Schedule a redundancy planning session to map appropriate patterns to your specific operations.
How do wastewater discharge permits affect greenfield timelines?
Wastewater discharge permits are among the longest-lead utility approvals: industrial pretreatment permits typically 6–12 months, NPDES direct discharge permits 12–24 months depending on receiving water and state. Permits gate construction in most jurisdictions because treatment system design depends on permit conditions, and treatment system construction is on the critical path. Start permit application 18–24 months before construction kick-off. Wastewater characterization (BOD, COD, TSS, specific industrial pollutants) drives both permit conditions and treatment system design — complete characterization in basic engineering, not detailed engineering.
What compressed air quality classes do we need?
Compressed air quality is specified per ISO 8573-1 across three dimensions: particles, water content (pressure dew point), and oil content. Typical specifications: Plant Air (general pneumatic): Class 3-4-4. Instrument Air (controls and pneumatic instrumentation): Class 2-3-2. Clean Dry Air or CDA (precision applications, sensitive measurements): Class 1-2-1 or better. Each quality class typically has separate generation, filtration, and distribution because higher quality classes require more expensive equipment that should be sized to the specific demand for that class. Compressed air typically costs $0.20–$0.40 per 1,000 SCF generated — leaks waste 20–30% of generation in untracked plants.
