The AI gold rush is hitting a wall. After months of unfettered spending on generative AI tools, major enterprises are now facing a harsh reality check. Uber reportedly blew through its entire annual AI budget in just a few months. Some companies have cut Claude licenses for parts of their organisation. Meta quietly killed its internal AI leaderboard. The era of "tokenmaxxing" — where CEOs encouraged employees to push AI usage as far as it would go — is giving way to a more sober assessment of return on investment.
The Tokenmaxxing Hangover
In early 2026, "tokenmaxxing" was the hottest trend in Silicon Valley. Companies encouraged employees to use AI tools for everything — from writing emails and generating code to creating marketing copy and analysing data. The logic was simple: if AI makes workers more productive, more usage must mean more productivity.
But the bills quickly mounted. Enterprise AI platforms like ChatGPT Enterprise, Claude Pro, and GitHub Copilot charge per seat or per token. Uber found that its annual AI budget was exhausted within months, leading to a company-wide clampdown. TechCrunch reported on the trend at the StrictlyVC Los Angeles conference: NEA's Tiffany Luck noted that "enterprises are still figuring out their AI ROI."
Who Is Cutting Back
| Company | Action | Reason |
|---|---|---|
| Uber | Exhausted annual AI budget in months | Uncontrolled token usage across teams |
| Meta | Killed internal AI leaderboard | Tokenmaxxing was not producing useful results |
| Several enterprises | Reduced Claude/ChatGPT licenses | Cost per seat exceeded productivity gains |
| Multiple firms | Halted experimental AI projects | Unable to demonstrate clear ROI |
What This Means for Indian IT and Startups
The AI budget pullback has significant implications for India's IT services industry. Indian IT giants like Infosys, TCS, and Wipro have invested heavily in building AI capabilities and training their workforce on generative AI tools. A slowdown in enterprise AI spending could mean reduced demand for AI consulting and implementation services.
On the other hand, Indian startups offering AI solutions with clear, measurable ROI — such as customer service automation, document processing, and code generation — may benefit as enterprises shift from experimental AI to practical, cost-effective applications.
Indian SaaS companies that can demonstrate a direct link between AI spending and business outcomes will be better positioned than those selling vague "AI-powered" promises.
The Path Forward: Practical AI Wins
The pullback does not mean AI is failing — it means the market is maturing. Early adopters who treated AI as a productivity magic bullet are now learning that effective AI deployment requires careful cost management, clear use cases, and measurable outcomes. Enterprises that succeed are those that focus on specific, high-value applications rather than blanket AI usage.
NEA's Tiffany Luck summed it up at StrictlyVC: the next phase of enterprise AI will be about "getting the ROI math right" rather than "maxxing tokens." For Indian companies, this means the opportunity is still massive — but only for those who can demonstrate real, quantifiable value.
