Ask anyone who has tried to build or connect a new energy project in the United States — it’s not just hard, it can be paralytic. Between infrastructure bottlenecks, skills shortages, and slow regulatory paths, the grid isn’t keeping up with demand. As people working on the software side of this longer-term evolution, we’ve come to appreciate something simple but often overlooked: we’re entering an energy abundance revolution, but our pipes — literal and digital — remain stuck in the past.
The U.S. electrical grid is now over a century old. It was built for a world where energy flowed in one direction, mostly from massive plants to homes and businesses. That world is over. What’s replacing it is far more dynamic and distributed — energy isn’t just being consumed at the edge, it’s now being generated and stored there too. Solar panels on rooftops. Batteries in basements. Microreactors for military bases and remote industrial sites. Even data centers — those new engines of economic growth — aren’t waiting for the grid to catch up. They’re building power on-site.
And for good reason. Interconnection delays now stretch up to a decade in some areas. Transformers are backlogged by more than 20 years. Transmission lines, where they exist, are loaded at just 50% capacity but often can’t handle new demand — either physically, politically, or economically. All of this is happening just as energy demand is spiking due to AI compute, EV adoption, manufacturing reshoring, and new consumer devices like electrical heat pumps.
The traditional model of energy — centralized, top-down, slow-moving — isn't much help in this new environment. What’s emerging is a more organic, fragmented grid: clusters of generating systems, batteries, software, and loads — each managed by different teams but ideally coordinated as a whole.
That’s where we believe software has a pivotal role. Coordination, especially at scale, is a problem tailor-made for digital intelligence. Today’s grid lacks visibility at the local distribution level. There’s no canonical dashboard for the energy internet — no Splunk or Looker equivalent for grid operators. In an industry where you can wait weeks just to get a drawing approved or even see load estimates, the software opportunity is both massive and underexplored.
AI adds another layer. We can use it not just to monitor and forecast but to actively streamline and accelerate regulatory reviews, simulate project impacts, validate permitting applications, and manage swarm-like clusters of distributed energy systems in real time. Complex it may be — but we’re long past the point where human oversight alone is sufficient.
As one example, Texas has been able to double solar capacity in the past few years. It wasn’t magic, just a combination of policy agility, entrepreneurial drive, and a little more software flexibility. We’d argue every state should follow that model — not just for solar, but for storage, grid services, and small modular reactors. These SMRs and microreactors, long an interesting but theoretical solution, now look increasingly practical. And not just because they’re clean or efficient, but because they shift generation closer to usage without requiring decades of new HV transmission.
Of course, building physical infrastructure still matters — a lot. We can’t software our way around the fact that most transformers are built using tech from the early 20th century, or that nearly every battery powering our EVs and backup systems is sourced from firms based in China. Rebuilding domestic capacity for transformers, batteries, and electric steel isn’t just an industrial policy play. It’s about resilience and, increasingly, national security. As one expert recently put it: “There is no national security without a stable electrical grid.”
Some people think the bottleneck is money. But we’re not short on capital — in fact, funding is flowing into the sector at record levels. The bottlenecks are space, skills, and speed. We’ve lost institutional memory on how to build large energy projects. The people who constructed our nuclear fleet have largely aged out, while skilled labor pivots toward tech or other industries. That expertise gap is a harder fix than it sounds.
That’s why we believe a truly resilient energy future will be portfolio-based. Solar, batteries, gas, wind, nuclear, and emerging technologies will all have roles to play. So will policy reform, workforce training, and smarter infrastructure planning. But the connective tissue — across all of it — will be software. Better tooling won’t just help manage demand; it will help us build faster, deploy smarter, and adapt continuously to localized needs, constraints, and shocks.
We’ve spent a lot of time thinking about what that software should look like. It won’t be monolithic. It will need to be modular, data-rich, and integrated with physical sensors and edge devices. It will need to adapt to regulation — not fight against it — and offer real transparency to both citizens and agencies. And most of all, it needs to be built with humility. The power grid isn’t just an engineering challenge. It’s a civic one.
There’s something sobering, but also inspiring, in the realization that the grid we rely on was largely built by people decades ago with tools we now consider basic. It didn’t scale because of exponential server capacity or machine learning. It scaled because people aligned behind a generational mission.
That same kind of systemic reinvestment is needed now — in infrastructure, in people, and in the tools that let us see and shape what’s next. Software alone can’t fix the grid. But without it, we’ll keep playing catch-up.