Harvey’s token – Harvey, a legal AI startup used by law firms and corporate legal teams, says its monthly AI token consumption has jumped from 1 trillion in January to an expected 12 trillion to 13 trillion for May. The CEO, Winston Weinberg, links the surge to the next busine
A legal AI startup that helps write, review, and analyze documents is burning through far more machine text than it was just a few months ago.
Harvey’s CEO. Winston Weinberg. said the company used 1 trillion tokens in January. speaking on the “Sourcery with Molly O’Shea” podcast. released on Tuesday. Tokens are the chunks of text that AI models process when users type prompts and receive answers—an accounting measure that can also shape billing.
Since then, Harvey’s monthly token use has ballooned roughly 12 times. Weinberg said the company was on pace to consume 12 trillion to 13 trillion tokens in a month. A Harvey spokesperson later confirmed that the estimate referred to May.
“The usage is getting insane,” Weinberg said.
Harvey’s product is designed for law firms and corporate legal teams, assisting with tasks such as drafting, reviewing, and analyzing documents. The CEO’s comments land as more tech companies start scrutinizing the cost of AI usage, especially when token-heavy systems produce large bills.
The industry shift is already visible in how some companies try to control spending. Business Insider recently reported that some startups are pulling back from “tokenmaxxing” after receiving high AI bills. Coinbase has been routing some prompts to cheaper models. and Uber executives have questioned whether spending on AI coding tools is translating into enough measurable product gains.
Weinberg said the next challenge for companies is choosing the right level of AI for each job. “You don’t want frontier intelligence running every task. It’s too expensive,” he said.
In a statement to Business Insider, Weinberg spelled out the logic with examples: “a change of control review warrants it. A first-pass document summary doesn’t.”
On the podcast, he compared the upcoming debate over AI spending to a familiar problem in law—how firms justify their bills to clients. He described it as the “billable hours problem.”
“When you get a bill from a law firm, it says in six-minute increments what people did, and then the hourly rate,” Weinberg said. “Why did they do that? It’s because they’re trying to show ROI.”
As companies spend more on tokens, he said they will face growing pressure to explain what that spending actually accomplishes.
“This is kind of like the main problem that I think the entire world is about to hit,” Weinberg said, adding: “Which is, ‘I just spent $1 billion on tokens. Where’s my ROI?’”
Harvey Winston Weinberg legal AI tokens artificial intelligence costs ROI billable hours problem May token usage law firms corporate legal teams tokenmaxxing Coinbase Uber AI billing
So basically lawyers are paying by the word now? Seems wild.
12 trillion tokens?? That sounds like a typo or marketing math. Like how is anyone even measuring that correctly if it’s “insane” lol.
Wait, I thought tokens were like crypto stuff, not text chunks. If they’re burning through tokens, doesn’t that mean the AI is just repeating the same documents more and more?
This is why I don’t trust “AI for law” at all. If the CEO says it’s too expensive to run the best model on every task, then what are law firms doing, just hoping nobody audits it? Also “billable hours problem” like… so it’ll still be charging clients for time even though it’s a machine.