By Priya Venkataraman
When a developer sets out to build a liquefied natural gas terminal, a carbon storage well, or a new solar farm, the public imagination tends to picture the hard parts as physical: the drilling, the construction, the grid connections. Inside the industry, a growing number of operators will tell a different story. The hardest, slowest, and often most expensive part of the project frequently happens long before anything is built. It happens in the permit file.
Permeta, an artificial intelligence company working in the energy sector, has built its entire business around that claim. Its message is blunt and easy to remember. “Faster permits. Smarter reviews. Powered by AI.”
A layer on top of the work teams already do
Permeta is careful to describe what it does not do. The company does not write permits, and it does not replace the geologists, engineers, and attorneys who assemble them. Instead, its software sits on top of the applications a developer’s team is already producing and reviews them for the issues that tend to trigger delay: missing sections, inconsistent figures, and citations to rules in the wrong form.
According to the company, that administrative back-and-forth accounts for about a third of a typical permit timeline across project types ranging from federally reviewed pipelines to carbon sequestration wells to wind and solar. The goal is to catch those problems before a regulator does, so the document that lands on an official’s desk is consistent, complete, and traceable from the first filing.
The company frames the benefit in terms regulators themselves describe. One state regulator quoted by Permeta put it plainly, saying that the single biggest improvement to the process would simply be submitting complete applications in the first place. Another noted that the repeated cycle of returning incomplete filings raises costs for developers, which ultimately works its way into the price of energy.
From a single well type to the whole sector
Permeta did not start with this broad mandate. The company launched in early 2025 as ClassVI.ai, named for the category of wells used to store carbon dioxide underground. Its original mission was narrow: streamline permitting for carbon capture and sequestration projects.
That focus did not last long, and not because it failed. As the team worked with developers, regulators, and policymakers, it became clear that the same friction existed almost everywhere in energy. A liquefied natural gas developer facing a complex federal pre-filing process had the same fundamental problem as a carbon storage operator. So did mining, wind, and solar projects. The company rebranded as Permeta to reflect a commitment to transforming energy permitting of every kind, from emerging energy sources to core infrastructure.
The early evidence is being watched closely. The company has said its first customer reported to the Houston Business Journal that its federal pre-filing process became up to 25% more efficient with Permeta’s help. The company describes its value in the same terms its customers use, with one permit writer calling Permeta the future of permitting and another customer saying it was the least painful permit they had ever done.
Why a general chatbot is not enough
A reasonable question hangs over any AI company working in a regulated field: why not just use a general-purpose tool that already exists? Permeta’s leadership has heard the objection directly, often phrased as some version of “I already have a chatbot, what do I need you for?”
The company’s answer rests on the stakes involved. A general model can produce text that reads well, but it can also hallucinate, leak sensitive data, and fail to explain how it reached a conclusion. In a permit application, where a single inconsistent number can send years of work back to the start, those weaknesses are not acceptable. Permeta emphasizes controls that keep client data protected, checks that hold the document to a consistent standard, and the ability to explain how the system arrived at a given result. Explainability, in particular, is something the company treats as essential rather than optional.
Selling against the status quo
Permeta’s most persistent competitor is not another piece of software. It is habit. The permitting process was designed in a world of paper and has changed remarkably little, even as software entered the picture. The company describes the typical workflow as serial and top-down, the kind of rigid, stage-by-stage planning that the software industry largely abandoned years ago in favor of faster, more iterative methods.
That makes adoption a cultural challenge as much as a technical one. Large organizations are slow to trust unproven technology, and some of the people closest to the work are openly wary of it. The company likes to point to a customer who began by saying she disliked technology and ended up unwilling to imagine doing another project without the tool. People who try it, the company argues, tend to change their minds.
Notably, Permeta declines to do what some technology vendors do when selling into a frustrated market. It does not blame regulators or treat them as the obstacle. The reviewers, the company stresses, are part of the system it wants to serve, not a target. The aim is to make the paperwork move, so that the experts on both sides can spend their time on the hard scientific questions that actually deserve it.
For an industry where a single delayed day on a major project can cost a million dollars, that is a pitch with real money behind it. The energy is in demand. The capital is committed. Permeta is wagering that the last great inefficiency left to solve is hiding in plain sight, in the stack of documents no one wanted to read.
