By Marcus Halloway
A mine operator once told Eric Redmond that a permit would take roughly three years to clear. Redmond asked a simple question in return: what would it be worth to cut that timeline to eighteen months? The operator did not hesitate. The answer was $450 million.
That number, and the certainty behind it, captures the problem Redmond has spent the last year building a company to solve. The founder and CEO of Permeta, an AI company focused on energy permitting, Redmond argues that one of the most expensive bottlenecks in building energy infrastructure has nothing to do with engineering, geology, or financing. It is paperwork.
“The world needs energy,” Redmond said. “Energy requires permits, and we can get them done faster with AI.”
A career built on systems
Redmond’s path to permitting was anything but direct. He graduated into the dot-com bust and, unable to find a job in a frozen market, spent a year playing professional poker while the industry recovered. He describes the experience less as a detour than as an education.
“Play the hand you’re dealt,” he said, summarizing what the game taught him. “Don’t complain about the cards you get, because your cards will come.” He has carried a related idea into business ever since. “Luck is just statistics taken personally,” he said. Over enough attempts, he believes, something is statistically likely to work. The trick is to keep playing well enough to still be at the table when it does.
When the job market thawed, Redmond moved into healthcare software during the early days of the Agile movement, an approach to building software that favors fast, iterative cycles over rigid top-down planning. From there he co-founded six startups, an experience he describes with unusual candor: two reached funding, two reached an exit, and two, in his words, he drove into the ground.
His systems work eventually brought him to Nike, where he spent roughly a decade and helped co-found the company’s Technology Innovation Office, a small team charged with bringing emerging technologies into one of the world’s largest brands. Along the way he wrote eight books, three of them bestsellers, including Deep Tech, a survey of the technologies he expected to define the decade. He placed artificial intelligence at the top of that list years before the current wave of attention arrived. He also completed an MBA at the Massachusetts Institute of Technology, adding a business credential to a background rooted in computer science.
Finding the bottleneck
The idea for Permeta surfaced after Redmond left Nike to work on tracking carbon credits in the energy sector. What he found was that the credits themselves were not the hard part. The hard part was getting the underlying projects permitted.
He started a company called ClassVI.ai, named for the class of wells used to store carbon underground. As he talked to developers, regulators, and operators, the scope of the problem widened in front of him. The same friction he saw in carbon sequestration showed up in liquefied natural gas, in pipelines reviewed by federal regulators, in mining, in wind, and in solar. The company rebranded as Permeta to reflect a broader mission across energy permitting of every kind.
The pattern Redmond kept finding was structural. Permit applications have grown so large and so complex that, in his words, they have become “beyond human knowability.” Highly educated teams of scientists, engineers, and attorneys assemble enormous stacks of documents, try to keep every figure and citation consistent, and inevitably miss something. The regulator catches the gap and sends the application back, triggering a request for additional information and another round of revisions.
“That ping-ponging can drive 30% of the timeline,” Redmond said. According to research he cites from McKinsey, major energy permits can take anywhere from three to ten years to clear. Thirty percent of that is measured in years, not weeks.
Building something people will use
Redmond is careful about how he frames the role of technology in a field that has run on paper for generations. He resists the temptation to blame regulators, a stance he holds firmly and repeats to his team. The goal, he says, is to help applications move faster, not to criticize the people reviewing them.
He is equally skeptical of the assumption that a general-purpose chatbot can do this work. A general model can produce a draft that looks polished, he argues, but in a high-stakes regulatory filing the user then spends just as long fixing the errors it introduced. Permeta’s pitch leans on control, consistency, and the ability to explain how the system reached a given result, qualities Redmond considers non-negotiable when a single inconsistent figure can send a multi-year application back to the start.
If there is a single principle Redmond returns to, it is one he repeats to his own team. “We don’t just want to build something useful,” he said. “We want to build something used.” A product can work flawlessly, he notes, and still fail if no one actually adopts it. That belief, more than any single feature, is what he says guides the company as it grows.
For an industry where every delayed day on a major project can cost a million dollars, the proposition is straightforward. The energy is needed. The projects are funded. The slowest step left, increasingly, is the stack of paper on a regulator’s desk. Redmond is betting that is exactly where the next decade of energy progress will be won or lost.
