Tarun Batchu has been chasing the same problem since he was 12. With Vela Energy, he's building AI agents for the unglamorous middle of the power grid - and getting projects moving.
Most 12-year-olds are memorizing state capitals. Tarun Batchu was building a carbon-capture project, earning a sponsorship from Arrow Electronics, and getting it in front of ExxonMobil. That single fact tells you almost everything about the way he works. He picks a problem that is far too big for the moment he's in, and then he refuses to put it down.
More than a decade later, the problem is still energy. But the angle has shifted in a way that only makes sense once you understand what he learned along the way. Batchu is now the co-founder and CEO of Vela Energy, a company building AI execution agents for large-load energy projects. Its mission fits on a bumper sticker: unlock the queue.
To understand why that mission matters, you have to understand the queue.
When a new power project - a data center load, a solar farm, an industrial plant - wants to connect to the grid, it doesn't just plug in. It enters an interconnection queue. It waits on engineering studies, procurement of long-lead equipment, permits across air, water, and zoning, and endless coordination between utilities, brokers, and regulators. This can take years. Not because anyone lacks the will to build, but because the execution work drowns the people doing it.
Batchu spent his teenage years learning this the slow way. He served as a UN Global Goals Ambassador. He was a national co-lead with Citizens' Climate Lobby. He organized community energy-efficiency initiatives backed by Bloomberg Philanthropies. That is a resume built on persuasion - convincing people that clean energy is worth wanting.
And that is the quiet pivot at the center of his story. Somewhere between the policy work and the code, Batchu stopped trying to change minds and started trying to change the plumbing. The belief was already there. What was missing was a way to move the paperwork.
Vela Energy's design is unusual for an AI company because it is deliberately restrained. The agents read the records a project already generates - schedules, drawings, permits, procurement data - across the tools teams already use, from Gmail and Google Drive to Autodesk and grid operators like ERCOT, PJM, CAISO, and MISO. No migration. They build one live model of what is actually happening, detect what changed across revisions, and trace the impact to milestones and the people who need to decide.
Then they draft a recommended action, with citations. And then they stop.
The company states the rule plainly: every conclusion names its source, and every write stops at a person. In an industry where a wrong move can mean a blown budget or a safety failure, that restraint isn't a limitation. It's the whole pitch. The engineers keep their authority. The agents just clear the runway.
Read Batchu's path start to finish and it can seem scattered. Carbon capture as a preteen. Climate policy and civic leadership through his teens. Now an EECS and business student in UC Berkeley's M.E.T. program, running a funded startup before he's finished the degree. Look closer and it's a single straight line drawn through one question: why is clean energy so slow to actually get built?
There's a useful lesson buried in that timeline for anyone building anything. Batchu didn't chase the exciting parts of the energy transition - the panels, the batteries, the headlines. He went looking for where projects actually stall, and he found it in the least glamorous place possible: procurement, permitting, and engineering studies. The boring middle. The part everyone wants to skip.
Batchu didn't build Vela Energy alone. His co-founder, Tony Li, serves as COO and brings a hardware-and-government pedigree that balances the CEO's policy-and-product instincts. Li worked on Tesla's thermal purchasing team, spent time at the U.S. Department of Energy, and studied at Vanderbilt. Between them, they cover the two sides of the same problem: the systems that move electrons and the systems that move decisions.
The company frames its values in the same plainspoken way it describes its product. Mission-driven: it exists to unlock the queue. Built on the ground: deployed where utilities and grid infrastructure actually operate. Human in the loop: engineers keep decision-making authority. Move first: fast execution without sacrificing diligence. It reads less like a manifesto and more like a set of rules a young founder wrote to keep himself honest.
When Batchu describes the goal, he doesn't reach for a modest one. Vela Energy, he says, is building the intelligence layer for the next generation of energy infrastructure. That is a big claim for a company this young. But it's consistent with the person making it - the same one who thought a middle-schooler could take carbon capture to ExxonMobil, and wasn't entirely wrong to try.
What makes the ambition feel earned is the restraint underneath it. Batchu isn't promising that AI will run the grid. He's promising something narrower and more useful: that the agents will do the drafting, the watching, and the tracing, and that a human engineer will always make the call. The technology moves fast. The accountability stays put.
If it works, the payoff isn't a flashy demo. It's a data center that connects a year sooner, a transformer sourced before the deadline, a permit that doesn't fall through the cracks. Unglamorous wins, stacked up. Which, if you've been paying attention to Tarun Batchu since he was 12, is exactly the kind of win he's always been chasing.