A new class of AI startups is bragging that they are taking money that would normally be used to hire people and spending it on AI compute instead. Astronomical AI bills are now, in a certain corner of the tech world, a supposed marker of growth and success.
Amos Bar-Joseph, CEO of Swan AI, a coding agent startup, wrote in a viral LinkedIn post recently that his AI bill just hit $113,000 in a single month for a four-person team. Bar-Joseph explained that his startup is spending money on Claude usage bills rather than on salaries for human beings and that the company is scaling with intelligence, not headcount.
The startup’s goal is $10 million annual recurring revenue with a sub-10 person organization. They do not have sales development representatives and their paid marketing budget is zero. However, they spend significant amounts on tokens. Bar-Joseph stated that the $113,000 bill is part of their go-to-market team, engineering, support and legal functions.
Writers have discussed tokenmaxxing extensively in recent weeks, describing it as a vanity metric at tech startups and major tech companies where spending on AI tools like Claude and ChatGPT signals productivity. The Information reported earlier this month on an internal Meta dashboard called Claudenomics, a leaderboard that tracks the number of AI tokens individual employees use.
The general narrative has been that the more AI tokens an employee uses, the more productive they are and the more innovative they must be in using AI. Stories abound of individual employees spending hundreds of thousands of dollars in AI compute by themselves. There has been at least a partial backlash with Salesforce inventing a metric called Agentic Work Units that attempts to quantify whether AI token spending translates into actual work.
Shifting so much money and attention to using AI tools is being done with the goal of replacing human workers. CEOs have justified mass layoffs with the idea that improving AI efficiency will reduce the need for human workers. Monday, Verizon CEO Dan Schulman said he expects AI to lead to mass unemployment.
While big companies are using AI to justify reducing worker headcount, startups are using AI to justify never hiring human workers in the first place. Chen Avnery, a cofounder of Fundable AI, stated that the $113,000 is not a cost but headcount budget allocated differently. Avnery runs a similar model processing loan documents that would normally require a team of 15 people.
Andrew Pignanelli, founder of General Intelligence Company, gave a presentation last month explaining that many of the jobs at his company are just a series of AI agents. He now usually spends more money on AI compute than on human salaries. Some days the company spends $4,000 on Claude Opus tokens.
Tokenmaxxing entrepreneurs often leave a key question unanswered: whether the spending on AI compute actually delivers value and remains financially sustainable. Companies like OpenAI and Anthropic are currently absorbing significant losses on their products because they price AI compute below its true cost.
