Why Sridhar Vembu is comparing AI dependency to drug addiction
18 Feb 2026

Sridhar Vembu, the co-founder of Zoho, has drawn a controversial parallel between artificial intelligence (AI) usage limits and drug addiction.

His comparison was sparked by a real-life incident at work where one of his team members struggled after running out of AI "tokens."

These tokens are essentially credits that power many generative AI tools.

The 'tokens exhausted' phenomenon
Token fatigue

Vembu revealed that a colleague of his complained about hitting a "tokens exhausted" wall.

The team member felt the AI was no longer as helpful as it used to be.

This wasn't about a software glitch, but rather the sudden loss of access to a productivity tool they had become reliant on.

Withdrawal symptoms?
Efficiency impact

Vembu compared the emotional response to losing access to AI tokens, with withdrawal symptoms. However, instead of chemicals, it was about losing efficiency. The slowdown in productivity, he implied, is similar to how people react when something they rely on is taken away without warning.

The productivity multiplier effect
Dependency concerns

Modern AI systems can perform tasks that once took hours in mere minutes.

This speed creates quick feedback loops, quick wins, and constant progress.

However, when users hit token limits and suddenly lose momentum, it feels worse than a normal slowdown.

Many professionals responding to Vembu's post echoed this sentiment, calling AI a "productivity multiplier" around which workflows are increasingly built.

Criticism of the analogy
Discussion

Not everyone agrees with Vembu's addiction analogy.

Some argue that the problem is not craving AI itself, but having work systems designed around constant AI availability.

When these tools disappear, productivity drops, much like losing access to cloud services or internal software platforms.

Others likened it to running out of paid credits in an arcade game: you know the limit exists, but the interruption still breaks the flow.

Over-automation and skill erosion concerns
Automation caution

Vembu's analogy also highlights a bigger concern of over-automation.

As AI becomes more integrated into daily work, there is a risk of delegating too much thinking, problem-solving, as well as creativity to machines.

If teams cannot operate smoothly without constant AI assistance, businesses might be sacrificing long-term skill resilience for short-term speed.

The ongoing debate in tech circles now focuses less on token pricing, and more on how dependent modern productivity is becoming on generative AI tools.