Startups Spend More on AI Compute Than Human Salaries: The Rise of 'Tokenmaxxing'

Startup CEOs are increasingly bragging about spending more on artificial intelligence (AI) compute than it would cost to hire human workers. In a viral LinkedIn post, Amos Bar-Joseph, CEO of Swan AI, revealed that his four-person startup spent $113,000 on AI tools in a single month, choosing to allocate funds toward AI usage rather than salaries.

Bar-Joseph emphasized that Swan AI is "scaling with intelligence, not headcount," aiming for $10 million in annual recurring revenue (ARR) with fewer than 10 employees. The company has no sales development representatives (SDRs) and no paid marketing budget, relying instead on AI-driven efficiency.

Bar-Joseph wrote:

"Our goal is $10M ARR [annual recurring revenue] with a sub-10 person org. We don’t have SDRs [sales development representatives], and our paid marketing budget is zero. But we do spend a sh*t ton on tokens. That $113K bill? A part of it IS our go-to-market team, our engineering, support, legal... you get the point."

What Is 'Tokenmaxxing'?

'Tokenmaxxing' is a term gaining traction in tech circles, referring to the practice of measuring productivity by the amount of money spent on AI tools like Claude and ChatGPT. The more an employee or company spends on AI tokens, the more innovative or productive they are assumed to be.

Recent reports highlight the growing obsession with AI token usage. The Information revealed that Meta tracks AI token consumption in an internal dashboard called "Claudenomics," ranking employees based on their AI tool usage. Some employees have reportedly spent hundreds of thousands of dollars on AI compute individually, setting a benchmark for others to aspire to.

Is AI Spending Translating to Real Productivity?

Critics argue that the focus on AI spending may be misplaced. Salesforce has introduced a metric called "Agentic Work Units" to assess whether AI token expenditures are generating tangible results. The shift toward AI tools is often justified as a means to replace human workers, with companies like Verizon citing AI efficiency as a driver for mass layoffs.

However, startups are taking this a step further by using AI to avoid hiring human employees altogether. Chen Avnery, cofounder of Fundable AI, commented on Bar-Joseph’s post:

"This is the part people miss about AI-native companies - the $113k is not a cost, it is your headcount budget allocated differently."

Avnery explained that his company processes loan documents using AI, a task that would typically require a team of 15 people. "The math works when your AI spend generates 10x the output of equivalent human cost," he noted. "The real unlock is compound scaling—token spend grows linearly while output grows exponentially."

Industry Reactions and Concerns

The trend has sparked debate over whether AI spending is a sustainable growth strategy or merely a vanity metric. While some startups and tech giants are doubling down on AI-driven efficiency, others are questioning whether this approach is scalable or even ethical.

Medvi, a GLP-1 telehealth company, is another example of an AI-native startup leveraging AI to minimize human labor. The company’s model relies heavily on AI processing to handle tasks that would otherwise require a larger workforce.