ServiceNow Customer Service Executive Warns ‘Tokenmaxxing’ Is A Cycle Of AI Hype

Tech workers are burning through massive amounts of AI computing in a race to code faster and automate more work. In Silicon Valley, this practice is known as “tokenmaxxing,” where employees push their use of ChatGPT and other large language models to the limit to increase productivity.

This trend has accelerated with the rise of AI coding tools and agents that can handle increasingly complex tasks. Unlike asking ChatGPT or Claude automatically, generating code and implementing an agent workflow consumes additional tokens, or data processing units for AI models when users enter information or make responses. Within startups and tech companies, the use of tokens is increasingly representative of how employees rely heavily on AI in their daily work, with some developers even leaving coding agents to work through the night to speed up product development.

But not every business AI leader believes that more usage automatically translates into better results. “I think this [tokenmaxxing] it’s going to be a short-term cycle,” Chris Bedi, chief customer officer at ServiceNow, told the Observer at the ServiceNow Knowledge 2026 event this week. “There’s a bill to pay for those tokens.”

ServiceNow is a large cloud platform that helps businesses manage, automate and design workflows. About 90 percent of Fortune 500 companies use its products, including Nvidia, AT&T and Delta Air Lines. In the first quarter of 2026, ServiceNow generated $3.67 billion in subscription revenue, a 19 percent year-over-year increase since making AI the centerpiece of its business strategy.

One of its flagship products is AI Control Tower, a platform that allows customers to oversee AI deployments, including tracking agent behavior and measuring return on investment (ROI). The company also offers “freelancers” working for “AI experts”, which it has nurtured to perform workflows across IT, customer relationship management, and security and risk, among other tasks. It is also investing heavily in developing AI capabilities through ServiceNow University, a platform designed to train employees to use AI in their jobs.

As AI agents take on larger roles in the workforce, Bedi says ServiceNow aims to help customers maximize value without overspending. But in his conversations with corporate clients, tokenmaxxing is not top of mind. “When I talk to the C-suite, tokenmaxxing doesn’t come up,” Bedi said.

Bedi argues that culture combines function and value. “It’s almost like measuring a restaurant based on how many ingredients it buys,” he said. “You don’t rate a restaurant that way. You can’t.” An employee who tells an AI chatbot multiple times to execute code, for example, can end up with the same result as a human who gets there with only a few instructions.

His skepticism comes as the use of AI within tech companies explodes. According to The New York Times, employees at AI companies use a staggering amount of computing internally, and one Anthropic employee allegedly made $150,000 in one month using Claude Code. Tokenmaxxing has become a symbol of how quickly the cost of AI testing can accelerate as operators face pressure to supercharge their workflows. As employers increasingly approve the adoption of AI in the workplace, the use of tokens is expected to increase significantly.

That move has been a boon for AI model providers who charge based on token usage. OpenAI claims its ChatGPT APIs process more than 15 billion tokens per minute. Google’s Gemini models now process more than 16 billion tokens per minute—a 60 percent increase year over year, according to its latest earnings report. For AI providers, business acquisitions create a powerful incentive structure: the more employees rely on AI, the more revenue those systems generate.

Some tech companies have encouraged employees to increase the use of tokens internally. At Meta, work created a token to track internal leaderboards and highlight top users, Information reported in April. After the project leaked publicly and sparked a debate about the value of the token’s use, the leaderboard was taken down.

Many token funds are increasingly treated like software fees or free food. At Nvidia’s annual GTC conference in March, CEO Jensen Huang said developers should expect an annual token budget worth about half of their already high salaries, on top of base salary, so that their output can be “increased by 10X.”

Still, a growing number of executives say tokenmaxxing risks becoming another metric of Silicon Valley vanity. Yamini Rangan, CEO of HubSpot, recently wrote on LinkedIn that “[Outcome maxxing >> token maxxing],” meaning that measurable business results are more important than implementation.” Andrew Lau, CEO of engineering intelligence firm Jellyfish, shared a similar sentiment, calling tokenmaxxing the “first place” to drive growth.

According to Bedi, the value of AI is best measured by whether it beneficially improves performance. Companies still rely on standard metrics: how much time they save workers, how much they produce, and whether AI improves efficiency.

Although much smaller than token calculations, traditional business results are still important in measuring AI ROI, Bedi said. “The ultimate goal is, how do I help my employees become as proficient as possible in AI, and how do I help them feel comfortable using it?”