May 29, 2026
As more companies buy into the AI “tokenmaxxing” craze, Michael Burry is becoming more bearish.
The famed “Big Short” investor recently suggested that tokenmaxxing — where companies try to maximize their AI tools based on the total tokens they process — is less about genuine demand and more about saving face in the heated artificial intelligence race.
Advertisement
As Business Insider reported, Burry described tokenmaxxing as a “crazy, rushed, temporary phase” that’s purely “quota-driven, leaderboard-driven, management-mandated overconsumption.”
According to Burry, tokenmaxxing is only happening because executives treat AI usage as proof of productivity and innovation. Supposedly, the more tokens a company blows through, the more it can tell investors it’s all in on AI, even if the actual business payoff is questionable at best.
These critiques are consistent with Burry’s prior comments published on CNBC, where he argued today’s AI boom is getting more and more similar to the late 1990s “dot-com bubble.”
One AI company that Burry believes will fall big time is mega-cap chipmaker Nvidia [NASDAQ:NVDA]. From a technical perspective, Burry claims trading activity for Nvidia is the “lowest since 1999,” which he thinks will make it harder to attract buyers on any signs of weakness.
And, judging by recent fundamental data, it wouldn’t take much in Burry’s opinion for Nvidia to falter. That’s because the bulk of its money is linked to just three clients. Business Insider reported that 64% of Nvidia’s accounts receivable come from three companies, up from 56% last quarter.
Advertisement
If there’s any shakiness in client demand, Burry believes Nvidia will feel it the worst, and that could be the domino that brings the data center driven rally down. Burry is also putting his money where his mouth is. His firm, Scion Asset Management, has purchased put options on 1,000,000 Nvidia shares.
The economics of AI hype
Whether sustainable or not, tokenmaxxing is a real obsession for some of the most powerful companies in the world.
For instance, Business Insider reports that JP Morgan and Disney now have “leaderboards” specifically for tokenmaxxing stats. Visa also said it upped its AI usage from one trillion tokens this February to two trillion one month later.
But for Big Tech companies, two trillion tokens is child’s play. According to Google’s CEO Sundar Pichai, monthly AI token processing is now at a mind-boggling 3.2 quadrillion tokens, roughly seven times higher than a year earlier.
The theory is that higher token consumption means employees must be more productive since they’re using AI more aggressively. But for critics like Burry, the number of tokens companies use is more about status and internal rankings than an actual revenue booster.
So, is there any data to back up Burry’s claim? With so many big companies spending so much on AI tokens, they can’t all be delusional, right?
Must Read
- You can now build wealth like a landlord for as little as $100 — and no, you don't have to chase down rent or take 3 A.M tenant calls
- Goldman Sachs used to hoard prime real estate deals for the ultrarich. Two ex-analysts just opened the door for $250
- Robert Kiyosaki begs investors not to miss this ‘explosion’ — says this 1 asset will surge 400% in a year
Join 250,000+ readers and get Moneywise’s best stories and exclusive interviews first — clear insights curated and delivered weekly. Subscribe now.
More tokens may not mean more dollars
It’s not just bearish investors that are questioning the tokenmaxxing trend, as even a few members of Silicon Valley’s upper echelon aren’t entirely sold on spending more and more on AI tokens.
Advertisement
For example, Uber’s COO Andrew Macdonald said he isn’t seeing the “link” between spending more on tokens and better profitability. As Macdonald said in a Business Insider report, “I think maybe implicitly there is more that is getting shipped, but it’s very hard to draw a line between one of those stats and, ‘Okay, now we’re actually producing 25% more useful consumer features.’”
Some preliminary data backs up these more critical comments on tokenmaxxing. For instance, the engineering analytics platform Jellyfish analyzed data from 12,000 software developers in Q1 2026 and found that the heaviest AI users consumed about 10 times more AI tokens than average users, but they only produced roughly twice as much work.
Granted, that’s a higher output, but it’s hard to justify the costs. Jellyfish noted that a typical developer spends about $52 on AI tokens per month. As for heavy tokenmaxxing users, they were burning through roughly $691 per month in AI costs alone.
Even more striking, the cost per completed code contribution jumped from just 28 cents for low AI users to about $89 for the heaviest users.
What all this suggests is that companies are paying dramatically more for a relatively modest productivity boost. Whether that will lead to an AI market apocalypse is debatable, but tokenmaxxing may cool off as companies take a fresh look at their bottom lines.
You May Also Like
- Dave Ramsey warns nearly 50% of Americans are making 1 big Social Security mistake — here’s what it is and the simple steps to fix it ASAP
- Thanks to Jeff Bezos, you can now become a landlord for as little as $100 — and no, you don't have to deal with tenants or fix freezers. Here's how
- Millionaires under 43 are reshaping investing — just 25% of their portfolios are in stocks. Here’s where their money is going
- Robert Kiyosaki issues grim warning for baby boomers. Many could be ‘wiped out’ and homeless ‘all over’ the country. How to protect yourself now
Eric Esposito is a freelance contributor on MoneyWise who loves making financial topics accessible and understandable to readers. In addition to MoneyWise, Eric’s work can be found in publications such as WallStreetZen and CoinDesk.
