Dan Shipper, the CEO of Every, a company spanning AI media, software, and consulting, recently revealed a startling personal expenditure: a $13,000 bill for OpenAI's Codex in a single month. This significant sum, which Shipper humorously noted earned him “side eye” from COO Brandon Gell, underscores the rapidly escalating costs associated with advanced AI adoption across the business landscape.
AI as a Core Business Expense
Shipper's personal use of OpenAI's Codex extends to automating various administrative tasks, including writing emails, navigating his schedule to suggest meeting times, and drafting replies. While the AI agent requires human approval before sending communications, its deep integration into Shipper's workflow highlights a broader organizational philosophy at Every.
The company treats AI tools not as an optional add-on but as a fundamental business expense, akin to providing employees with laptops or health insurance. Every employee gains access to AI resources, with technical staff receiving higher-tier plans that can cost around $200 per month. Crucially, if employees exceed their allocated AI usage limits, Every covers the additional “token budgets,” normalizing these costs within operational overhead.
The Rise of 'Tokenmaxxing' and AI Spending
Every's approach reflects a growing trend in the tech industry, sometimes referred to as “tokenmaxxing,” where companies invest heavily in AI services to unlock productivity gains. This phenomenon is not isolated to startups; recent reports indicate that some consulting firms have spent hundreds of millions of dollars on AI services, such as Anthropic’s Claude, within a month. Such figures highlight increasing concerns over the soaring financial implications of integrating AI at scale.
Shipper emphasizes the transformative potential of AI, noting that it can empower employees to operate like managers earlier in their careers. “Many, many more people are capable of that than we think,” he stated. However, he also acknowledges that AI tools, despite their sophistication, still require human oversight and intervention to be effective and responsible.