Sponsored content by Bernadette Johnson | General Manager and Head of Power & Renewables, Enverus

I recently spoke with a solar developer who told me land was no longer the difficult part of the process. She could find a thousand acres with good irradiance and a willing landowner in an afternoon. What she couldn’t find, at least not without months of work, was the answer to the only question that mattered. Would the grid let her build there?

That gap, between where solar can technically go and where it can actually connect, has become one of the defining challenges facing the industry. It’s worth being precise about why.

The reflex in 2026 is to assume AI has already solved it. It hasn’t. Most of what passes for AI-powered siting today just draws a more attractive version of the map developers already had. The harder, more valuable work is ranking sites by interconnection access, flagging basis risk before it erodes returns, and handing the survivors straight into engineering. That’s the challenge we set out to help developers solve.

The grid is the constraint, and it isn’t going away

Start with the data, because it frames everything that follows. In our 2026 Global Renewable Energy Trends Report, which surveyed a broad cross-section of energy professionals and included design data from more than 64,000 solar and storage projects, grid saturation and instability were the most-cited barrier to progress, named by 63.7% of respondents. Permitting and regulation came second at 47.8%.

Those numbers are striking. What matters more is that it has held at elevated levels for several years running. This isn’t a temporary backlog that better processing will clear. Congestion, curtailment, and delayed interconnection are becoming structural features of high-penetration markets. They’re reshaping where projects can viably be built, which technologies developers choose, and how investors price risk.

The interconnection queue tells the same story. More than two terawatts of generation and storage capacity are now waiting for a path to the grid in the United States alone, more than twice the country’s installed capacity. Most of that capacity will never be built; it’s withdrawn, fails its studies, or stalls for years. Reforms such as FERC Order 2023 are beginning to change these dynamics, and I’m genuinely optimistic about the direction they’re taking. But the lesson holds. A site chosen without a serious read on grid conditions is a site that may never produce a kilowatt-hour, however good the sun.

So the right question isn’t whether AI can help with solar siting. Of course it can. The question is whether a given AI system incorporates meaningful grid intelligence.

From a market thesis to a ranked site list, in minutes

Here’s what the work looks like when the grid is built into it from the start.

You begin with a thesis, not a parcel. A region, a technology, an offtake target, a return threshold. With Enverus ONE, our agentic workflow platform, that thesis becomes the starting point for a generation siting flow. It screens more than 156 million parcels, applies buildability filters, and ranks the survivors by the factors that actually decide a project’s fate: interconnection access and locational marginal pricing signals. The output isn’t a heat map for someone to interpret. It’s a ranked, auditable list of sites your team can act on.

What makes the ranking valuable is the market context underneath it. The workflow draws on decades of locational marginal pricing patterns, congestion events, and queue behavior. It surfaces grid-ready zones and flags basis risk at the individual parcel level. That’s the risk that never shows up on a satellite image but quietly decides whether a project’s economics survive its first year.

Here’s the distinction that matters. A general-purpose AI model can read a map and tell you where the land is flat, sunny, and unencumbered. It has no operating context for the grid. No proprietary history of how a node has priced power. No read on what the local queue is doing. No memory of the congestion events that will shape tomorrow’s curtailment. Enverus has spent 25 years accumulating exactly that context. Generic AI gives you a map. This gives you a defensible decision.

The value isn’t the technology. It’s weeks of analyst time compressed into an afternoon, and the confidence to walk into an investment meeting with a site list that holds up.

Narrowing the field before engineers ever open a file

A ranked list is a strong start. But it isn’t a buildable shortlist. At least not yet. The next step is rigorous exclusion, and it’s where a great deal of wasted effort gets cut.

Before any site advances, it runs through more than 50 layers of geospatial exclusion analysis within Enverus PRISM®.  Wildfire exposure. Proximity to substations. Queue status. Environmental and land-use constraints. Basis risk. Sites that survive that screening aren’t merely attractive on paper. They’ve already cleared the filters that tend to kill projects late in development, when sunk costs are highest, and a schedule slip hurts most.

Then comes the handoff that changes the economics of the whole process. The sites that survive screening move directly into design for full photovoltaic and storage engineering. No re-keying of data. No export and reimport. No starting the engineering conversation from a blank page. Your engineering team inherits a set of sites that have already been vetted against grid reality.

The discipline this enforces is the point. Engineers stop spending hours, sometimes weeks, on detailed designs for sites that were never going to clear interconnection screening. When only a fraction of queued projects ever reach commercial operation, the sequence you follow isn’t a matter of preference. It’s the difference between a team that compounds its effort and one that keeps designing projects that die in the queue. Screen first, then design. It’s one of the most underrated levers in development.

Better inputs, better designs, more financeable projects

When pre-vetted, grid-aware sites feed the design engine, every downstream decision improves. Our platform within RatedPower produces layouts, equipment selections, energy yield estimates, and the detailed engineering and financial outputs developers use to demonstrate bankability. A connected workflow provides analysis, design details, and documentation that highlight your project’s bankability, without leaving investors with doubts or open questions. Because the sites arrived already screened for interconnection access and basis risk, the design work is grounded in grid reality from the first iteration. You’re not reassessing whether a project is viable halfway through engineering. You’re refining one you already trust.

This is where the value of the sequence compounds. Better siting produces better designs. Better designs produce projects that clear financing with fewer surprises. The direction of travel is clear in the data. Hybrid solar-plus-storage projects are on the rise. Storage is now central to protecting project economics in saturated, volatile markets. The ability to design solar and storage together, on sites already proven grid-ready, has stopped being a nice-to-have. It’s the core competency.

None of this removes engineering judgment from the process. It does the opposite. It frees skilled engineers to spend their time on the projects most likely to be built, not the ones that look good until the grid says otherwise.

Why the connected workflow matters

Step back, and the advantage is the connection itself. Plenty of tools do one piece of this. What I haven’t seen anyone else do is move from AI-powered siting across 156 million parcels that draws from over 25 years of proprietary data, through rigorous grid and exclusion screening, into a complete engineering package, without you ever leaving the platform or rebuilding your data along the way. Every handoff that disappears, therefore, removes a source of error, delay, and lost context.

That matters because of where we started. Grid congestion is the number one barrier to solar development, cited by nearly two-thirds of professionals, and it has stayed there for years. A faster map doesn’t address it. What addresses it is grid intelligence applied at parcel scale, grounded in proprietary energy data no general-purpose model can replicate, and wired directly into the engineering decisions that follow. That’s the work Enverus ONE was built to do.

I’m optimistic about where this is heading. The constraint is real and structural. It’s also, in the truest sense, solvable. The developers most likely to succeed over the next several years will be those who treat the grid as the first question rather than the last, and who put AI to work where it actually changes the answer.

About the Author

Bernadette Johnson joined Enverus in 2016 through the acquisition of products and services from Ponderosa Advisors, where she was a founding partner. Throughout her career in energy, Bernadette has earned a reputation as an industry expert with extensive experience providing crude, natural gas, and NGL fundamentals analysis and advisory services to various players in North American and global energy markets. In 2022, Bernadette was promoted to GM of Power & Renewables, where she continues to lead Enverus Power & Renewables division. She holds an M.S. in International Political Economy of Resources and a B.S. in Economics from the Colorado School of Mines.