Amidst doomsday predictions of a job apocalypse due to a surge in automation, two Silicon Valley companies have shown signs of distress, calling AI more expensive. Companies like Microsoft and Uber have noticed that deploying human workers can be economical.
Microsoft has reportedly begun cancelling most of its direct Claude Code licenses. According to The Verge it is instead moving engineers toward using GitHub and Copilot CLI. This comes just six months after the firm first opened up access to Claude Code, encouraging thousands of its developers, project managers, designers, and other employees to experiment with coding.
However, it’s worth noting that cancelling Claude Code licenses won’t affect Microsoft’s Foundry deal. Under which it’s investing up to $5 billion in Anthropic and giving Foundry customers access to Claude models, as well as Anthropic’s $30 billion commitment to purchase Azure compute capacity, reports The Verge.
Uber is scaling back automation investments
Microsoft isn’t the only company scaling back its internal AI use. Uber’s COO in a recent interaction, warned that the AI costs are rising fast, but there aren’t many productivity gains yet.
‘That link’s not there, Andrew Macdonald told the Rapid Response podcast.
Earlier, Uber’s CTO Praveen Neppalli Naga suggested that the firm had already burnt through its entire 2026 AI coding tools budget in just four months. This comes after the company had actively incentivised adoption through internal leaderboards ranking teams by AI tool usage.
These developments also suggest that the cost of replacing human labour with AI may be more complicated than some predictions.
“While AI can deliver meaningful productivity gains, the total cost of ownership, spanning compute infrastructure, model inference, integration, governance, and human oversight, remains substantial. In many enterprise use cases today, AI is serving more as a productivity enhancer and workforce augmentation layer, rather than a direct and economically viable replacement for human workers,” Prabhu Ram, Vice President – Industry Research Group, CMR told Financial Express Digital.
Agentic AI to transform job landscape
Goldman Sachs also recently forecasted that agentic AI could drive a 24-fold increase in token consumption by 2030 as consumers and enterprises adopt AI agents. It will reach a staggering 120 quadrillion tokens per month. As businesses turn to AI agents to boost productivity, aggregate costs could rise sharply even if the price of each token falls.
But as consumption increases, the cost of individual AI tokens is expected to fall sharply. A recent report from research firm Gartner found that by 2030, a highly sophisticated AI model will cost AI firms nearly 90% less than it did in 2025.
Even so, Gartner predicted that cheaper tokens won’t translate to cheaper enterprise AI because agentic models require far more tokens per task than standard models, increased consumption can outpace falling unit costs, and AI providers won’t fully pass through lower costs to consumers.
“Chief Product Officers (CPOs) should not confuse the deflation of commodity tokens with the democratisation of frontier reasoning,” Gartner senior director analyst Will Sommer warned in a statement.
Tech giant’s AI obsession
These cautious calls come at a time when tech majors are engaged in an emphatic race to succeed in the AI one-upmanship. A Meta employee crafted a leaderboard, fittingly named “Claudeonomics,” after Anthropic’s AI model, to track which workers are using the most AI.
Amazon is pushing its employees to “tokenmaxx,” or use as many AI tokens as possible (the basic building blocks of AI compute). But with a token-based pricing system, the work gets more expensive with more use and better efficiency.
“Yes, in many cases, it comes down to token usage. Simpler consumer queries use fewer tokens and therefore require lower compute resources. However, enterprise and corporate use-cases often involve significantly larger and more complex queries, including long documents, data analysis, coding, reasoning, etc,” Abhilash Kumar, Lead Research Advisor (Director) at Smart Analytics Global noted.
According to Kumar, as token usage increases, the AI model requires more GPU compute and electricity to process the request and generate high-quality responses. This increases inference costs for AI companies, especially for Microsoft.