Storage development: from pipeline to connection

00:00Owen Feld: Good afternoon, everyone. Our next session is from Dana Whitcomb, who runs development at Cedar Grid. Dana has been building storage projects for about twelve years, and she has agreed to talk about the part of the process most of us complain about in the hallway. Dana.

00:31Dana Whitcomb: Thanks, Owen. I want to start with what our pipeline looks like, and then spend most of the time on what happens to a project after we announce it, because that is where the schedule is really set.

Our project pipeline is growing. Announcing a project is only the first step; getting it connected is what turns a plan into an operating asset.

Today Cedar Grid has 23 storage projects, about 3.1 gigawatt-hours in total. Two are operating. Four are under construction. The other 17 are in development, which for us means we have site control and a queue position, and in most cases not much else yet. Two years ago the pipeline was 14 projects and about 1.6 gigawatt-hours, so it has roughly doubled.

03:15When people see a pipeline number like that, they tend to read it as capacity that is coming. Some of it will come. Some of it will move, and some of it will not get built. We announced four projects for 2024. Two reached operation on the date we first gave. One moved by a year and is under construction now. One was withdrawn after its second cluster study came back with upgrade costs we could not carry.

So when I say pipeline, I mean a list of plans. Before a project becomes an operating asset, it still has to clear connection approval and then equipment delivery. Those are two different gates, and different people run them.

05:40The other change is size. Our early projects were 50 megawatts with two hours of storage. Most of what we file now is 100 to 200 megawatts with four hours. A larger project is not much harder to permit, but it can be much harder to connect, because it asks more of the substation and the lines around it.

07:20One slide on how we stage spending, because it explains how we think about risk. We spend little until the first study results arrive, more after the second, and we commit to long-lead equipment around the time the agreement is signed. That staging used to follow the battery order. Now it follows the queue.

08:42Dana Whitcomb: Which brings me to the queue.

Battery availability has improved for our projects, but the connection queue still determines the schedule. A ready battery does not mean a ready grid connection.

Two years ago, if you had asked me what kept me up at night, I would have said cells. We were placing battery orders a year ahead and still worrying about allocation. Today our enclosure suppliers are quoting about five months, and prices are lower than they were. That is good news, and it has improved the economics of a few projects that did not work before.

It has not changed the schedule of a single one.

11:05Our last six interconnection agreements took an average of 35 months from request to signature. The fastest was 26 months. The slowest was 44. For most of that time there is nothing we can do but answer study questions and post deposits. We have one project that has been through two restudies because other projects in its cluster withdrew, and every withdrawal moved the upgrade costs around.

13:30I want to be careful here, because the queue gets blamed for everything. The operators are handling a very large number of requests, some of them speculative, and the rules have been tightened to deal with that. Site control requirements and higher deposits have helped. But the volume of storage requests is still rising, and the studies take a long time because the network questions are hard. Battery availability improved; the grid connection still sets the pace.

15:50What we changed internally: we now build every project schedule backward from our best estimate of the agreement date. We used to build forward from site control. That one change made our dates more honest. It also changed what we announce. We now give a target year only after the first cluster study results, not at filing.

18:10People ask whether we should file more requests to improve the odds. We did some of that a few years ago. It costs money in deposits, it adds to the load on the study process, and in our experience it does not make any single project faster. We would rather have fewer projects at better sites.

19:45One more point on batteries. Cheaper cells help. They let us size projects for four hours where two hours used to be the limit. But a bigger battery behind the same connection point does not reach the grid any sooner.

21:10Dana Whitcomb: The last thing I want to cover is what happens after approval.

Connection approval is our main constraint today. On some projects, transformer delivery becomes the next critical path after approval.

When we sign an agreement, the clock moves to equipment. The longest item for us is the main power transformer at the project substation. Quotes we received this year ran from 12 to 18 months. High-voltage breakers are next, at about a year. The battery enclosures, as I said, are about five months.

24:30So on a project where we wait for the agreement and then place the order, the transformer can set the energization date. On two of our projects this year, the transformer arrived after the date we had planned to energize. We lost about four months on one and two months on the other. That is where most of our equipment delays now come from.

27:05Our response has been to order main transformers earlier, at agreement signature or sometimes before it, with a cancellation fee if the project fails. That is a real cost, and we only do it where the study results give us confidence. We have also standardized on two transformer ratings, so a unit ordered for one project can move to another.

30:20To sum up: the pipeline is growing, battery supply is in good shape, connection approval sets most of our schedules, and transformer delivery is next in line. Thank you.

30:58Owen Feld: Thank you, Dana. We have time for a few questions.

31:20Audience member: You said cheaper batteries have not changed any schedule. Do you expect that to change if queue reforms speed up the studies?

Dana Whitcomb: If the studies get faster, the transformer becomes the constraint on more of our projects. That is the honest answer. Faster approvals would move the bottleneck, not remove it.

34:40Audience member: How do you decide which announced projects to drop?

Dana Whitcomb: Mostly the study results decide for us. If the network upgrade costs come back above what the project can carry, we withdraw at the next decision point rather than wait and hope.

36:05Audience member: Have transformer prices come down the way cell prices have?

Dana Whitcomb: No. Our transformer quotes went up a little this year, and suppliers now ask for larger deposits to hold a production slot. A few manufacturers have announced new factory lines, and we are glad to see it, but that capacity is more than a year away. I do not plan around it yet. For now we plan around the quotes we have in hand, and we ask for a delivery window in writing before we commit to an energization date.

38:15Owen Feld: Last question, and it is mine. What would you tell someone reading a pipeline chart?

Dana Whitcomb: Read it as plans. Ask how many of the projects have a signed agreement, and how many have their long-lead equipment on order. Those two numbers will tell you more about next year than the total.

40:10Audience member: By that test, where does your own pipeline stand?

Dana Whitcomb: Of the 17 projects in development, three have signed agreements and are in procurement. The other 14 are still in study. We have main transformers on order for two of the three, and we expect to order the third this quarter. So most of our pipeline is still waiting on the queue, and a small part of it is waiting on equipment. If you asked me the same question two years ago, the answer would have been batteries for almost all of it. That is the change I wanted to describe today.

42:48Owen Feld: Thank you, Dana.