What does “tokenmaxxing” mean? Tech bro's newest obsession

In recent times, many technology companies are increasingly pushing their employees to adopt artificial intelligence tools in their daily work. Organizations such as Visa and JPMorgan, among many others, along with AI-sector companies themselves (including Anthropic), have introduced a range of incentives to encourage the adoption of these tools within their teams.

Behind these initiatives lies the idea that the more employees use artificial intelligence systems, the greater their productivity. This is the so-called "tokenmaxxing" — a term built on the "-maxxing" suffix, widely used in social media language to indicate the extreme optimization of a given behavior. In this context, "tokens" are the units of text processed by AI models — words, sentence fragments, or pieces of code; the expression therefore describes the tendency to delegate an ever-growing number of tasks to generative tools, maximizing their use in daily work.

What "tokenmaxxing" consists of

One of the reasons behind the spread of "tokenmaxxing" is the growing importance of AI agents. Unlike traditional chatbots, which respond to specific user requests, these systems can carry out complex tasks over extended periods — including developing an application, designing a website, analyzing large amounts of data, or coordinating articulated tasks — all without human intervention. In practice, the user assigns a task and lets the agent work for several consecutive hours.

This approach, however, radically changes the scale of computational resources required. Classic interactions with a chatbot — such as asking it to revise a document or summarize an email — consume a relatively modest volume of tokens. Agents, on the other hand, produce and process enormous amounts of information, instructions, and code during each operational phase. According to some estimates, a single agent continuously engaged on a project can "burn through" hundreds of millions of tokens in the course of a single week.

In some tech companies, the push to use these tools more and more has even fueled internal competition among employees, contributing to a rapid rise in the infrastructure costs associated with AI. The website The Information reported the case of a Meta programmer who allegedly used around 280 billion tokens in a single month — a volume that, according to various estimates, may have cost the company nearly one and a half million dollars.

Why high token consumption cannot be sustainable

For many companies, at least at this stage, the cost of intensive AI tool usage is considered a strategic investment. The goal is to accelerate the internal adoption of these tools and push employees to integrate them into every phase of their work. However, high token consumption carries a significant environmental impact: AI systems require enormous amounts of electricity to process large quantities of information, as well as large volumes of water needed to cool data centers.

Despite this, in some cases the number of tokens consumed by teams is even publicly displayed on social media as a badge of innovation. Startups and large financial groups share leaderboards, internal statistics, and usage records to demonstrate how heavily their employees rely on automation. More than a simple operational metric, token consumption is becoming an identity signal: a way to present oneself as a forward-thinking company, capable of rapidly adopting new technologies and turning them into a competitive advantage.

Not everyone in the industry, however, is convinced that this AI race is sustainable in the long run. Within the very companies that offer near-unlimited access to artificial intelligence models, doubts are beginning to emerge — particularly regarding the actual usefulness of investments that can reach thousands of dollars per day for a single employee. The critical point concerns above all the relationship between costs and benefits: increasing the number of tokens consumed does not automatically mean working better, nor does it guarantee that investments will truly translate into greater productivity or tangible economic results.