In February this year, a group of large customers found out what their power actually cost.
Twenty-four applicants had asked Salt River Project for 7.2 gigawatts of new power. For scale, SRP’s whole system peaked around 9 gigawatts this summer, a level it took the utility more than a century to reach. SRP worked out what the grid upgrades to serve those applications would cost and asked for deposits against that number.
According to an SRP presentation to the Pinal County Board of Supervisors in August, most of the applicants dropped out when the first deposit came due.
Those projects were in the queue. They had land, or the promise of it. Some had engineering money allocated. What they did not have was a price, and the moment they got one, the project stopped being real.
This is the normal pattern, not an Arizona anomaly. Lawrence Berkeley National Laboratory tracks interconnection data covering nearly all of the country’s generating capacity, and found that of the capacity that requested interconnection between 2000 and 2019, only about 13 percent had actually reached operation by the end of 2024. Roughly three quarters of it was withdrawn. For projects that did get built in 2025, the median time from first request to operating was over five years.
So before you commission transmission line design, the useful question is not how long the design will take. It is whether the project survives contact with its own cost.
Five questions get you most of the way to that answer.
1. Has the utility told you what your load triggers, and what it costs?
This is the one that kills projects, so it goes first.
A request for power is not a request for a wire. Depending on what you are asking for and where, serving it might mean a new line, a new substation, an upgrade to equipment miles away from you, or all three. The utility works this out through a study, and until that study runs, nobody knows the number. Not you, not your engineer, not the utility.
Arizona utilities have formalised this. SRP introduced a process in 2025 that requires a study specifically to identify what infrastructure a new large customer triggers, and gives that customer a cost estimate for their own upgrades. SRP also now applies minimum billing to new accounts above 20 megawatts of forecast load, charging the greater of actual demand or 80 percent of what was forecast. That second rule exists because customers were overstating what they needed.
Proceed if you have a study result and a cost estimate in writing.
Get more information if you have a study underway. Do not commission full design against an assumption about what it will say.
Stop if nobody has started the study. You are not ready to design. You are ready to apply.
2. Do you control the route, or only the endpoints?
Knowing where a line starts and ends is not the same as knowing where it can go.
A transmission or distribution line needs a continuous strip of land it has the legal right to occupy. Every parcel along that strip is a separate negotiation, and any one owner can force a reroute. Add in state highway crossings, railroads, canals, federal land, and tribal land, each with its own approving authority and its own timeline, and the route becomes a legal problem before it becomes an engineering one.
Land acquisition also does not run on your schedule. It runs on the schedule of whoever owns the land, and some of them will not answer the phone.
Proceed if the route is controlled or the acquisition is substantially complete.
Get more information if you have a preferred route and no agreements. Price the alternates now, while changing them is free.
Stop if the route crosses something you have not identified the approving authority for. Find out who that is first.
3. Where is the voltage number coming from?
There is usually a moment early in a project where someone states a voltage. Sometimes it comes from the utility’s study. Sometimes it comes from what the last project used.
Those are very different levels of confidence, and the difference shows up later, because voltage drives almost everything else. Structure type and height. Easement width. Clearance requirements. Equipment cost and lead time. Which permits apply. Whether the project needs a substation at each end or can tie into what exists.
Design produced against the wrong voltage is not partially useful. It is redone.
Proceed if the voltage comes from the utility’s study or an executed agreement.
Get more information if it came from a planning assumption. Confirm before drawings start.
Stop if two documents in your own project file disagree about it.
4. Does the equipment fit inside the schedule?
Transformers, switchgear, breakers, and large conductor have all had long order times since 2021, and those have not fully recovered. Demand from data centre construction across the country is competing for the same manufacturing capacity.
The failure mode here is specific and it repeats. A project builds its schedule around the design and permitting sequence, which is the part it can see, and orders equipment once the design is final, which is the part it cannot see. The design finishes on time. The steel arrives eleven months later.
Work it the other way. Take the date the project needs power, subtract the longest lead item, and see whether the design and approval work fits in what remains. If it does not, the answer is to order long lead items against a preliminary design, accept some risk, and say so out loud.
Proceed if long lead items are identified and ordered or scheduled against the energisation date.
Get more information if you have a schedule but nobody has checked current lead times this quarter. They move.
Stop if the energisation date came from a lease or a financing document and was never tested against procurement reality.
5. Who pays if the number comes back higher?
Uncomfortable, and worth asking before the design rather than during construction.
Utility upgrade costs in Arizona are increasingly assigned to the customer driving the growth. Both APS and SRP have said publicly that large new customers pay the cost of serving them, which is deliberate policy to keep those costs off residential bills. Reasonable policy. It also means the number in the study is your number, not the utility’s.
If the study returns higher than the project modelled, somebody absorbs it. Establish who before anyone has spent money, because after the spend the conversation gets harder and the relationships get worse.
Proceed if there is a written position on cost overrun and contingency.
Get more information if there is a budget but no contingency for utility-assigned upgrades.
Stop if the project economics only work at the low end of the estimate range.
What this looks like in Arizona right now
Some context on why these questions are sharper here than in most states.
At an Arizona Corporation Commission workshop in April, the state’s three largest electric utilities described roughly 8 to 10 gigawatts of speculative large load projects sitting in their queues. APS has said the Phoenix metro area ranks second in North America for proposed data centre development, and that the average data centre in its territory draws about as much power as 64,000 homes.
Most of that will not be built. That is not a criticism of anyone. It is what a queue looks like when applying is cheap and building is not.
The practical consequence for anyone doing real work here is that utility study queues, engineering capacity, and equipment supply are all being consumed partly by projects that will withdraw. Being early and being specific are worth more than they were three years ago.
A note on what counts as transmission
Worth clearing up, because it changes who you need and what the project involves.
Higher voltage lines moving power across long distances are transmission. Lower voltage lines distributing power to customers are distribution. In between sits subtransmission, which does a bit of both, and which is where a great deal of practical utility work actually happens.
The distinction matters because the permitting, structure design, clearance rules, and approval path differ across those categories, and a firm that works fluently in one is not automatically the right fit for another. When you scope the work, name the voltage rather than the word. “69 kV overhead” tells an engineer more than “transmission” does.
How ARUSI approaches this
ARUSI is a Service-Disabled Veteran-Owned Small Business (SDVOSB) engineering firm in Phoenix. We have designed utility infrastructure across Arizona since 1987, which means we have watched several of these cycles arrive and leave.
Our work on line projects is engineering and coordination. We do not build. On a project at this stage, that usually means route study and alternatives, structure and clearance design, coordination with the serving utility’s standards and study output, right of way and permitting support, and the drawing production that turns all of it into something a contractor can price and build.
The reason we would rather talk before the design is commissioned is that the five questions above are cheap to answer and expensive to skip. A project that fails question one does not need drawings yet. Telling a client that costs us a design fee and saves them considerably more.
Talk to us before you commission the design. Bring the utility study if you have one. If you do not have one, that is the conversation.
