Req 25329DD · Segment Director/Advisor, Power Markets — Utilities Built for the Enverus product team · August 2026

The grid has a throughput problem. Enverus is holding the best cards in the industry. I'd love to help play them.

I'm AJ — an electrical engineer who spent sixteen years on the buying side of energy software: control rooms, switchyards, and lately the PJM interconnection queue. This page is my read on the utility opportunity and a few ideas I'd want to pressure-test with your team. I built it myself, over a few evenings, because honestly this is my idea of fun.

Andrew J. "AJ" Lambert BSEE · MBA · MS Finance · Reading, PA

Years in power — plant,
field, and RTO

0GW

Through PJM's reformed
process on my watch⁷

0GW

In PJM Cycle 1 —
the first reformed cycle⁸

0GW

Waiting in US queues
nationwide¹

The story below

01 · About me

Four rooms that shaped how I think

Pick a room. Each one taught me something I still use every day — and together they explain why I want to build software for this industry rather than keep buying it.

The build

PJM Interconnection · Manager, Interconnection Projects

Since 2023 I've led the team executing PJM's FERC-mandated interconnection reform — issuing the studies and agreements for the Fast Lane and Transition Cycle 1, and serving as PJM's lead for dispute resolution and FERC filings.¹¹ Along the way I got to do the two most fun things in my career: originate NextGen, PJM's interconnection platform, and carry it from a problem statement to a live product¹² — and serve as the business lead for PJM's collaboration with Google's Tapestry, learning first-hand how an RTO evaluates, bounds and governs AI inside a FERC-jurisdictional process.¹³

And here's the honest part: those projects are why I'm writing to you. PJM is a great institution and I'm proud of the work — but software is a means to an end there. Operating the grid, planning the future, running markets: that's PJM's core competency, as it should be. I found the part of my job I never want to stop doing, and it's the product part. I want to do it at a company where software is the product.

People firstThe best thing about this industry is the people in it — linemen, dispatchers, developers, regulators. Most of what I know, someone in a hard hat taught me.

Puzzle addictI chase hard problems for fun. The interconnection queue is one of the best puzzles in American infrastructure — that's half of why I never left it.

Energy nerdI find this industry endlessly interesting: physics, markets, policy and human nature all colliding in real time, with real stakes.

Builder at heartI'm learning software and product development as a hobby — this site is a product of that. Give me a blank page and a gnarly workflow, and I'm happy.

02 · What I see in the market

The demand is real. The queue is jammed. And paper isn't power.

Three forces are colliding, and the third one is the opportunity almost nobody is serving. Every number links to its public source at the bottom of the page.

Force one — demand growth nobody planned for

Five-year US peak load growth forecast, by edition · GW — hover the points

Data centres broke the forecast: 24 to 166 GW in three editions — a 7× revision.² NERC now projects +224 GW of summer peak over ten years, 69% above its own prior-year view.³ Every planning assumption in the industry is being rewritten at once — and every rewritten assumption needs new tooling.

Force two — a queue drowning in its own volume

The crunch is technical AND administrative

Waiting now

2,061 GW

Roughly twice the country's installed capacity, queued behind cluster studies and restudies.¹

Arriving yearly

~2,300

New requests per year — applications that can run thousands of pages each, reviewed by human engineers.¹

On the clock

Firm deadlines

Order 2023 replaced "reasonable efforts" with hard study deadlines and penalties.⁹ The paperwork now has a compliance clock on it.

The technical work (network models, cluster studies, cost allocation) and the administrative work (intake, documents, milestones, agreements) are both past human scale — and they bottleneck each other. I've run this process. Fixing one without the other fixes nothing.

Force three — the one that keeps me up at night: surviving the queue still doesn't get steel in the ground

US requests 2000–2020, by capacity · status at end of 2025 — hover the segments

75% withdrew

13% operating 75% withdrew ~10% waiting

And the number that matters most: even among projects that survive everything and sign an interconnection agreement, 41% of that capacity still never gets built.¹ The median project that did make it took 61 months from request to operation.

The part I don't think anyone has figured out

A signed interconnection agreement is permission — it is not power. A stack of studies proving the grid won't overload if something gets built doesn't serve a single megawatt of load. Between "agreement executed" and "megawatts flowing" sits the least-served stretch of the entire pipeline: network upgrades, equipment with multi-year lead times, contractor schedules, financing milestones, energisation — coordinated across a developer, a transmission owner and an RTO who each see only their own piece.

The country needs real energy, quickly, and the tooling for this last mile barely exists. In my view, whoever builds it doesn't just win a product category — they become the system the build-out runs on. That's a real need with revenue to match, and it's wide open.

03 · The map

Where the pieces sit today

The interconnection software market, one square per RTO seat. Tap any cell — this is the picture I keep coming back to, and the reason I got excited enough to build this page.

Market	Enverus SUGAR	GridUnity workflow	Google Tapestry	Nira dev-side
PJM144 GW active
MISO382 GW active
SPP171 GW active
CAISO191 GW active
ISO-NE15 GW active
ERCOT408 GW active
NYISO29 GW active

Swipe the map sideways · tap a cell

Tap a cell to explore

What jumps out at me

Enverus already holds the hardest thing in this market to build: MISO benchmarked SUGAR at 99%+ agreement with its incumbent tool, and a sitting FERC Commissioner cited it by name. Two seats held, two completely unclaimed, and the process around every study keeps growing. The rest of this map is the conversation I'm hoping to have.

Enverus / SUGAR GridUnity workflow Tapestry (in production) Developer-side only Unclaimed

What this map says to me

The MISO benchmark is the technical high-water mark of this entire market — a public, reproducible validation no competitor can match. Two seats held, two completely unclaimed, and around every study engagement the process itself keeps growing: intake, milestones, restudies, agreements, and eventually construction. Where the next layer of value sits in that stack is a question I have a strong view on from the buyer's side of the table — and it's the first thing I'd want to whiteboard with the team, because it deserves a conversation, not a webpage.

A brief word on the rest of the portfolio: I know Enverus also brings design tooling used across most of North America's largest utilities, a governed AI platform, and a research team already shaping the load-growth debate. The people who built those know them far better than I ever could from outside — so I'll skip the lecture and just say the foundation looks real, and the market above is bigger than any of us.

04 · The machine

The grid is the largest machine humans have ever built. Every piece of it is an opportunity.

Follow the power — generator to meter. At every stage there's equipment to design, build, operate, maintain and replace; people who do that work; and software those people mostly tolerate rather than love. Tap each stage: I've worked most of them in person.

Swipe the diagram sideways · tap a stage

Generation · Where the electrons start

Generation

The equipment

Turbines, reactors, wind, solar, batteries — plus the step-up transformer that puts it all on the grid. Thousands of plants, each with a maintenance program, an outage schedule and a control room.

The people

Plant engineers, control-room operators, outage planners. I held a Senior Reactor Operator licence — I know how these teams think, and exactly what they will and won't trust.

The tools today

Work-management systems designed for compliance, not for the person using them. Paper procedures. Decades of historian data nobody can easily ask a question of.

The opening

Predictive maintenance, outage optimisation, procedure intelligence — AI that shows its work, because an operator only acts on what they can verify. Enormous, untouched, and buyable as a service.

Utilities aren't built to innovate — and that's the opportunity

Design it. Build it. Operate it. Maintain it. Replace it. Five verbs, forty-year assets — and at every verb, software the people using it mostly tolerate. I've watched a lineman demonstrate the app that adds twenty minutes to his day; he wasn't wrong. Better data for the engineer is necessary but not sufficient — the tool has to fit the way the work actually flows, so the people doing it want to reach for it. That flip is the entire game.

Now the structural part, said with respect: innovation isn't a utility's business, and it shouldn't be. Their business is spending money well to make money reliably — deploy capital prudently, earn the regulated return, keep the lights on. Software R&D doesn't fit that model and never will. So give them the version that does fit: a partner whose products let them invest in reliability and lower cost — outcomes they can carry to their commission and, better yet, brag about to their customers, which is a utility's favourite thing to do. The utility wins with its regulator and its ratepayers. The partner earns a seat that lasts decades. Everybody gets to say so out loud. That's the business.

05 · The field

Who's already selling to utilities — and why the door is still open

An honest scan of the landscape. Every corner has vendors; what nobody owns is the workflow layer that ties a utility's day together. The market isn't waiting to be taken from a rival — it's waiting to be created.

Who	What they sell	The read
The operations incumbents GE Vernova · Hitachi Energy · Siemens Grid Software · AspenTech OSI · OATI	The control room: EMS, SCADA, ADMS, market management systems.	Deeply entrenched and genuinely hard to displace — but hardware-led, slow-cycle, and built for the control room rather than the field. Displacing them takes a decade. Going around them takes a quarter — the growth is in workflows they've never touched.
The engineering desktops Siemens PSS®E · GE PSLF · PowerGEM TARA · ETAP · PSCAD · Bentley · AutoCAD	The tools every planning and design engineer opens each morning.	Per-seat, single-case, file-based — respected and stuck. The workflow between these tools is email and shared drives; I've lived it. That connective tissue is completely unowned, and it's where the hours actually go.
The data & analytics crowd Yes Energy · Wood Mackenzie · Aurora · Ascend Analytics · Modo · Amperon	Market data, price forecasting, resource planning.	Strong franchises on the market-participant side of the house. Much thinner on the wires side — the asset, field and process data utilities actually own and can't use. Different buyer, different sale, mostly different companies.
The AI point solutions AiDash (vegetation) · Pano AI (wildfire) · Buzz Solutions (inspection) · Utilidata (grid edge) · Neara (network models)	One sharp problem each — venture-backed and growing fast.	Proof the appetite is real: AiDash alone counts 135+ utility customers. But each sells a point answer, so the utility ends up integrating a dozen vendors itself — exactly the burden it was trying to shed. The whole-product seat is empty.
The interconnection specialists	Study engines and queue workflow — covered in the map above.	The most regulator-encouraged corner of the market, and the one where Enverus already holds the public benchmark. The beachhead is established; the question is what it grows into.

Swipe the table sideways

How big, how fast

The spend is real and accelerating

AI ops spend '25

$6.2B

AI ops spend '26

$7.6B

Utility-operations AI spending is growing ~22% a year²⁹, riding on $1.4T of utility capex through 2030⁶. Surveys put 81% of North American electric utilities already using AI somewhere — but analysts describe the buying as aggressive and "not always strategically planned," with mixed returns.³⁰

What that pattern means

Utilities are buying AI without a trusted guide

Eighty-one percent adoption with mixed returns isn't a maturing market — it's a market buying ahead of its own understanding. Every one of those disappointing pilots makes the next purchase more cautious, and makes the vendor who actually delivers more valuable.

So the share worth chasing isn't a slice of some existing category. It's the guide's seat: the partner who knows the work from the inside, picks problems worth solving, and delivers tools people reach for. Whoever earns that seat first ends up defining a category the analysts haven't named yet.

Where I'd attack first, knowing what the buyer knows

Attack one

The last mile: agreement to energisation

The build gap from Force three. Three buyers feel it — developer, transmission owner, RTO — no vendor serves it, and the revenue starts the day the pain does, which is today. This is where my queue experience converts most directly into product.

Attack two

Asset intelligence where the supply chain bites

Transformer lead times measured in years turn asset-health data into money. The records already exist — test results, maintenance history, loading — scattered and unused. I spent two years at PPL generating exactly this data; I know where it lives and why nobody can query it.

Attack three

Large-load screening for the data-centre wave

Every utility is improvising its answer to hundred-megawatt requests. A structured screening and readiness product meets the single loudest pain in the industry — and it's the fastest credibility play, because the demand shock is on every executive's board agenda already.

Attack four

Field tools crews actually want

Cheapest to prototype, fastest to prove, and the most underestimated moat in the sector — because adoption loyalty compounds and rips out slowly. This is the purest fail-fast play on the list, and the one where watching the work beats any spec sheet.

06 · Ideas I'd love to pressure-test

Three product bets, with my honest doubts attached

These aren't recommendations — I don't know your roadmap, your costs, or what's already in flight. They're the first ideas I'd bring to a whiteboard, each with the reason it might work and the part I'm least sure about. Tap to open.

Why it might work

Most queue pain enters at the front door: incomplete applications, shaky site control, model errors. A tool that lets a developer self-check against tariff criteria before filing prevents bad inputs instead of processing them faster — the highest-leverage point in the whole pipeline. Enverus already has the developer customer base, and every RTO would quietly cheer.

What I'm not sure about

Whether developers will pay for it directly or whether it works better bundled into PRISM as retention. And each RTO's criteria differ enough that the content layer is real work — I'd want to size that honestly before promising seven markets.

Why it might work

I've sat on calls where a TO receives a network upgrade scope and starts re-keying it into their own systems by hand. One chain — study result → upgrade scope → SBS design → construction tracking — would serve the most underserved buyer in this market — and few companies are better placed to offer it end to end.

What I'm not sure about

TO procurement is slow and committee-heavy; the champion is usually a planning manager without budget authority. I'd want to find two design-partner utilities before building anything — and I have relationships that could get those conversations started respectfully.

Why it might work

Utilities buying AI face prudency review with no established standard to point at — the hardest conversation in the sale happens inside the buyer's own building. I've assembled that evidence from the buyer's side: oversight design, audit trails, the documentation a regulator actually asks for. Understand the barriers well enough to design for them from day one, and the approval process flips from obstacle to differentiator.

What I'm not sure about

Whether it's a standalone product or a design principle baked into every product. Possibly the latter — which would make it the cheapest idea on this list, and the most durable.

And a tease There are more where these came from — including one about Europe, where the same queue crisis is unfolding at national scale with no incumbent vendor in sight. But I'd want to earn a real view on that market before pitching one. Consider it teased.

07 · How I'd work

Find a problem. Build fast. Put it in real hands. Earn the next one.

The cost of trying has collapsed — AI-assisted building turns the two-quarter project into the two-week prototype. This page is a small proof: a few evenings, my own research, my own code. The discipline isn't avoiding failure; it's picking problems worth failing at, and failing fast enough that it's cheap.

Step one

Start where it hurts

Ride along. Sit in the dispatch room, walk the yard, watch the planner reconcile two models by hand. The problem you find in the field beats the one you assume in a conference room — and the person who showed it to you becomes your first user and your loudest champion.

Step two

Build small, build fast — use AI to build AI

Modern tooling means a working version in weeks, not roadmap-quarters. Ship the smallest thing that touches the real workflow. If it's wrong, you've spent weeks learning why — that's not failure, that's the cheapest market research in the industry.

Step three

Iterate until they want it

Put it in the hands of a lineman, a planner, an operator — and watch. If they route around it, that's data, not disobedience. The bar isn't "adopted after training." The bar is reached for without being told. Iterate until you clear it.

Step four

Expand what earns trust

One real workflow, working, buys the next conversation. Relationships compound faster than features in this industry — utilities are careful with new partners and loyal to proven ones. Land, deliver, grow. Win-win or don't bother.

The regulatory barriers are design inputs, not legal review

Every product in this space lives inside a rulebook — FERC processes, NERC reliability standards, state prudency reviews. Most vendors treat that as a compliance checklist at the end. I've spent years on the inside of those barriers: filing at FERC, answering the committees, assembling the evidence a regulated buyer needs before they're allowed to say yes. Understand the barriers deeply enough, and you can design products to clear them from day one — which is exactly what utilities need and cannot build for themselves. That's where domain knowledge plus AI tooling stops being a pitch and starts being a moat.

08 · One possible path

A five-year shape — a sketch to react to, not a plan

If I were lucky enough to work on this, here's roughly how I'd think about sequencing. The team would make every part of this better; that's what teams are for.

Listen, learn, earn one win

Foundation

I'd start by earning the right to have opinions: the SUGAR deployment teams, the SBS and RatedPower account folks, and thirty customer conversations using relationships I already have. Then aim small and real:

1.1

One bounded new engagement at an RTO

Not a moonshot — a scoped piece of work that produces a public reference. The regulator is actively encouraging exactly this¹⁹, which makes the timing unusually kind.

1.2

Ship the pre-validation tool to the developer base

Small, fast, and it makes every future RTO conversation easier — "we're already improving what enters your queue."

1.3

Show up where the standards are being written

EPRI's AI consortium and the critical-infrastructure AI conversations now forming at the federal level are open to participants today.²³ Influence there is cheap now and unbuyable later.

Let the wins compound

The pivotal year

Year two is about following the customer's pull. Every study engagement raises adjacent asks — the trick is saying yes to the right ones, deliberately.

2.1

Grow each seat into the process around it

Restudies, milestone tracking, agreement support, construction visibility — the asks arrive on their own once you're in the room. Which to take, and in what order, is a portfolio decision I'd want to make with the team, not for it.

2.2

Open the unclaimed seats

ERCOT (408 GW, no incumbent, no Order 2023 hook — a genuinely different sale) and NYISO. Uncontested ground is the cheapest ground.

2.3

Pilot the TO workbench with two design partners

Find two transmission owners willing to co-build the study-to-design chain. My old world; I'd love to make those calls.

Make trust the moat

Differentiation

By year three the market will be crowded with AI claims. The durable position is being the vendor whose numbers regulators and engineers trust.

3.1

Turn regulatory fluency into product

The barriers — FERC process, NERC standards, prudency review — are design inputs. Ship products that arrive with the oversight design and evidence a regulated buyer needs, because that's what utilities want and can't build alone. First vendor to make that a habit defines what "trustworthy" means in this market.

3.2

Be ready for the CIP cloud window

The cloud pathway for CIP-scoped systems opens around 2029.²⁶ Designing against the drafts now means arriving on day one while others start reading.

3.3

Co-publish the benchmark

With a consortium or lab partner, so it reads as science rather than marketing.

Go where the same crisis lives

Expansion

The interconnection crunch is global, and most of it is unserved. With the US map won, the playbook travels:

4.1

Keep one eye on the same crisis abroad

Britain is triaging a 738 GW queue through a brand-new gating process; Germany rewrote its connection rules this spring; no vendor owns either.²⁵ I won't pretend expertise on markets I haven't worked — but a playbook proven at home travels, and I'd want us positioned to be asked when those doors open.

4.2

Canada as the natural first step out

Alberta has received data-centre applications exceeding its entire provincial peak. Familiar rules, familiar engineering culture — expansion that doesn't require becoming a different company.

4.3

Then the platform story

Application → study → agreement → design → build → energise: every step already lives somewhere in the portfolio. Joined up over five years, that's not a bundle of products — it's the system of record for how new generation gets built.

A plan you can't falsify is a pitch. Here's what would prove mine wrong, and what I'd watch for:

If a workflow player bolts on a credible engine — or gets acquired by someone with one — the account layer closes. That risk is why I'd force the build-or-buy question in year two, not year four.

Utility-owned consortium models could erode any pure-model advantage. Fine — compete on workflow, data and trust instead, which were always the defensible layers. Being inside the consortium beats watching it.

Forecasters themselves warn data-centre demand may be overstated. But the thesis doesn't need new demand: 2,061 GW is already in the queue, and clearing the existing backlog is a decade of work by itself.

They might — RTOs build a lot. The counter is the things internal teams rarely sustain: cross-market benchmarking, a support organisation, and contractual accountability. And a vendor who genuinely understands RTO governance gets invited in rather than resisted — that's a relationship skill more than a product one.

09 · Say hi

What I'd actually bring to the team

I'm not going to claim I know your business better than you do — I don't, and this page will have things wrong in ways your team will spot in minutes. What I can offer is the other half of the picture: sixteen years as the engineer, the operator, the field supervisor and the RTO manager this industry's products are built for, plus two years learning what it takes to originate and ship software inside its most regulated corner.

And one thing worth saying plainly: I'll deliver in the seat as scoped — but I think about this market at the segment level. Where the revenue is, what to build first, how trust compounds into share, what the team looks like when it works. The demand shock has opened a window that won't stay open, and I want to spend it building something that matters, with people who move fast. However far you want to take that ambition, I'm in.

Discovery that starts warmI can sit with a planning engineer, a dispatcher or a VP of transmission and hear the problem underneath the requirement — because I've held most of those jobs, and this industry runs on trust between people who've done the work.

Credibility in hard roomsRTO committees, FERC processes, TO engineering reviews — the rooms where deals in this segment actually live. I know their rhythms, their vetoes, and their people.

Builder's energyI love this work. I taught myself enough product and software to originate a platform at PJM, and enough to build this page from scratch. Give me forty hours a week of it and I'll be the happiest person on the team.

If any of this sparked a thought — even a "here's where you're wrong" — I'd love the conversation. Those are usually the best ones.

ajl5102@gmail.com (610) 406-8345 LinkedIn Back to top ↑

Andrew J. Lambert · Reading, PA · BS Electrical Engineering, Penn State · MBA, Villanova · MS Finance, Penn State · Former NRC Senior Reactor Operator

Source register