This startup CEO spent $13,000 AI bill on OpenAI’s Codex in a single month

In recent days, the term “tokenmaxxing” has been spamming our social and news feeds, as companies spend millions of dollars on AI services for productivity gains. Recently, a consulting company reported spending over $500 million in 30 days on Anthropic’s Claude, showcasing concerns over the soaring costs of AI adoption. Now, Dan Shipper, CEO of Every, a company operating across AI media, software, and consulting, revealed that he spent around $13,000 on personal Codex overage charges, according to a Business Insider interview.

Must read: AI spending nightmare: Companies spend over $500 million in 30 days on Anthropic’s Claude

The report further revealed that Shipper recalled the spend as the highest AI bill he has paid. He said, “I got some side eye from our COO, Brandon Gell, on that one.”  In addition, his response to a year-over-year comparison was simple: ” Way more. Way, way, way, way, way, way, way, more.”

Shipper further revealed that it heavily relies on OpenAI’s Codex to write emails on his behalf. He lets the coding agent go through his emails, look at his schedule, suggest suitable times for meetings, and also write draft replies for him. However, it can not send emails without approvals.

Despite the massive AI spending, the company is treating the spend as a normal business expense rather than an add-on. Shipper highlighted that every employee gets access to AI tools as part of their job, and they also get early access to OpenAI’s and Anthropic’s AI models. Its full-time employees use the basic $20/month AI subscription, whereas technical employees get higher-tier plans costing about $200/month.

Must read: Why AI agents are costing companies more than the employees they replaced

He further revealed that if employees exceed their AI usage limits, the company pays the extra costs, and highlighted that these “token budgets” are treated like standard employee costs, similar to things like laptops or health insurance. Reportedly, Every has integrated AI deeply into its workflows, which offers reasonable gains. Shipper highlighted AI tools and an agent that helps its employees work like managers, earlier in their careers.

Shipper said, “Very few people actually get the opportunity to be managers, because managing humans is very expensive and risky. I think many, many more people are capable of that than we think.” However, he also acknowledged that AI has its limits, and it requires human intervention.