Uber Takes the Leap: 30,000 Employees with ChatGPT Work

Uber announced this week the rollout of ChatGPT Work across its entire global workforce — approximately 30,000 people. The expansion follows a pilot launched earlier in 2026 with 1,000 employees and represents a significant milestone in enterprise AI adoption: this is no longer an experiment, but operational infrastructure.

From pilot to infrastructure

ChatGPT Work was launched in July 2026 as an enterprise-focused agent platform. Unlike the standard ChatGPT, Work integrates directly into the daily tools teams use — Google Drive, Gmail, Google Calendar, and Slack — enabling employees to perform complex, multi-step tasks and deliver completed work, from data analysis to meeting briefs.

Notably, of the early adopters in the initial pilot, about 80% were non-engineers, according to Sumanth Sukumar, Uber's VP of engineering. The finance, operations, and product departments were the first to adopt the tool, with use cases including analysis of complex operational data to identify trends, localization of training materials for different markets, and building personalized dashboards for daily management.

"Our approach is to make powerful AI tools available across Uber while giving teams room to experiment," Sukumar said. "Earlier this year, we paired 30 AI engineers with teams across Uber and gave them two weeks to automate key parts of their job. Those 'AI pods' showed us how much opportunity there is when you combine technical expertise with deep understanding of a specific job."

The engine: GPT-6.1 Sol and cost optimization

A detail that has received less attention is the model choice: ChatGPT Work will run by default on GPT-6.1 Sol, a cheaper model released by OpenAI in late September. Uber frames this choice as a deliberate balance between performance and cost efficiency — and it makes perfect sense in the context of the company's AI strategy.

Uber has already demonstrated a better understanding of AI cost challenges than most tech companies. In March 2026, the company burned through its annual AI budget in just three months. Since then, engineering teams have implemented dozens of optimizations: optimized model selection, weekly benchmarking, cheaper subagents for smaller tasks, automatic prompt compaction above 400K tokens, and, most significantly, a transition to open-weight models that cost 2 to 20 times less than frontier models.

The result, as documented by Uber itself, was a 34% reduction in cost per AI request and a 52% reduction in cost per session, while usage continued to grow significantly. Total cost has remained flat since March despite substantially more tokens being processed.

The bigger picture: AI, efficiency, and workforce reduction

The ChatGPT Work rollout arrives at a moment of transition. In July 2026, Uber laid off 10% of its customer service team, explicitly citing AI adoption as the reason. In June, it cut 23% of its people operations department. CEO Dara Khosrowshahi has been consistent in positioning AI as an operational imperative, not a luxury.

But the ChatGPT Work rollout differs from direct layoffs. It's about equipping remaining employees to do more with less — which, in practice, may mean the same outcome, but with a different narrative. The company did not disclose how much it's spending on the rollout, and Sukumar said Uber will "track how employees are using the technology" to guide future AI decisions.

OpenAI, meanwhile, gains a powerful reference case. With direct competition from Anthropic in the enterprise market, and Anthropic itself preparing for an IPO with a target valuation of $2 trillion, each large corporate deployment serves as a calling card. The fact that Uber had used Claude Code as its preferred tool earlier in the year and is now making ChatGPT Work the standard reflects the reality of the enterprise market: customers aren't loyal to brands, they're sensitive to cost-performance.

What to watch

The question is no longer whether AI will transform knowledge work — the question is which companies will balance performance, cost, and human adoption before their competitors do. Uber has taken a significant step in that direction, but the real test will come in the coming quarters when the company begins publishing concrete productivity and adoption metrics.

Sources: TechCrunch, OpenAI, The Pragmatic Engineer, Bloomberg

✓ Independent sources cross-checked and verified before publishing