Samir Kaul is fed up with high legal fees. The managing director of Silicon Valley's Khosla Ventures says he has been pushing back on lawyer bills. In one instance, Kaul told a private equity firm investing in one of his portfolio companies that he only refused one term in its proposed agreement: the budget for $1.5 million in legal fees. The total should be a tenth of that amount, Kaul recalls saying.
When Kaul started in venture capital in the early 2000s, an early round investment term sheet known as a Series A cost $25,000. Now, he said, it's $200,000. "What should it cost me?" he said. "It should be, like, a couple hours of an agent's time." An agent, not a lawyer, Kaul said.
Major investors and asset managers like Khosla have powered the dramatic growth of traditional law firms for the past decade or more. They are the types of clients Big Law can't afford to lose. Now, they have a new way to push back on legal bills: AI will do the work.
Norm Ai's rise
Some are taking that thesis one step further, placing bets on companies that will use artificial intelligence to redesign legal services. The financial incentive to work with those firms makes this round of law firm startups different from prior versions that fizzled out.
Norm Ai, an East Coast-based tech company, has raised $260 million from Khosla, Blackstone Inc., Bain Capital, and others. It was valued at $1.2 billion in July. The investors are already clients of Norm and its law firm, Norm Law.
"My team uses them," said Kaul, who has used Norm for work on fund formations, closing simple venture transactions, and forming special purpose vehicles for follow-up investments.
Norm got its start concentrating on a narrow band of compliance and transactional work that is well-suited for automation. On two full floors of Manhattan's 1 World Trade Center, roughly 50 lawyers are developing AI agents. The attorneys are mostly Big Law refugees whose experience includes stints at firms such as Cravath, Swaine & Moore, Skadden, Arps, Slate, Meagher & Flom, and Wachtell, Lipton, Rosen & Katz. The systems they're designing are meant to streamline a never-ending flow of compliance work - unsexy tasks similar to copy-editing for highly regulated clients.
"In a traditional law firm, they're leverage-and here they're skilled," Norm AI co-founder John Nay said of the lawyers.
Nay never worked at a law firm. He's an AI researcher who sold an investment platform last year to Nuveen. Norm's genesis came in a 2022 academic paper in which Nay warned against trying to train AI on ethics. He argued instead that encoding the law into AI is the best way to align it with democratic values. The paper also briefly mentioned that if AI understood the law, it could help improve the practical activity of law. Nay launched Norm in 2023.
Big Law takes notice
Big Law ears pricked up when Blackstone poured significant funding into Norm in November. That's when Norm launched its law firm, noting that Blackstone was a partner to "shape and develop" Norm Law. Blackstone has paid Kirkland & Ellis, on average, around $70 million in annual legal fees for the last five years. Lawyers from Kirkland and Simpson Thacher & Bartlett have helped Blackstone put billions to work. They declined comment.
Nay said on a podcast this month that Norm is doing routine Big Law work. But he clarified he isn't using AI for "bet-the-company" matters - legal industry lingo for difficult problems. Blackstone said publicly in May that Norm Law's work for the firm on an investment matter was handled at a lower price than traditional processes. Norm has since handled more work and continues.
Norm Law's January hire of Michael Schulte, former chief foreman at Hinckley & Allen, added a measure of credibility. The law firm side of the business is doing work that is "human dependent" and would otherwise be done by the firm's attorneys. Norm Law also has ties to competitors. Jeff Hammes, who built Kirkland into the country's largest firm, is an investor.
Norm Law's partner roster now totals 18 lawyers, all former corporate attorneys. They work side-by-side with software and legal engineers. The goal is to bake the partners' knowledge into repeatable AI workflows. A company spokesperson declined to provide revenue. Khosla led the latest funding, raising $120 million in July.
The dynamic extends beyond Norm. Matt Souza, co-founder of rival firm Talairis Law Group, framed the central question. "The real test, for the field, is whether and how that moves into the messier, judgment-heavy parts of the law," he said. "That's the open question."
For legal professionals watching this space, the implications are direct. The pressure on fees isn't coming from client negotiations alone anymore - it's coming from investors who are funding alternatives. Firms that automate routine work, whether through in-house AI development or partnerships with startups, will be better positioned to retain clients who now have a credible alternative. The lawyers most exposed are those doing repeatable transactional and compliance work; the lawyers least exposed are those whose value rests on judgment and high-stakes strategy that AI can't yet replicate.
Your membership also unlocks:
