English 
5 min of reading 
0 comments 
Uber consumed in four months the entire 2026 budget for artificial intelligence and adopted cost control in tools used by developers. The case exposes the doubt about AI productivity, governance, and financial return, as companies try to balance innovation, expenses, and large-scale corporate adoption.
Artificial intelligence entered a phase of cost control within Uber after the company consumed, in just four months, the entire 2026 budget allocated for the area. The decision also exposes a growing doubt about AI productivity in companies, according to a report published by Exame on June 13, 2026, based on data released by Bloomberg and confirmed by the company.
The measure mainly affects tools used by developers, such as Claude Code and Cursor, and establishes a monthly limit of $1,500 per employee for each AI-based programming software. If the spending exceeds this cap, it will be necessary to provide justification and obtain internal approval.
Uber sets a limit where there was acceleration before
Uber’s decision draws attention because it comes at a time when companies from different sectors are rushing to adopt artificial intelligence in their routines. The technology has moved from being just a promise to occupying areas such as software development, legal, marketing, and customer service.
-
The 2014 World Cup left a million-dollar legacy in Itaquera: the Corinthians stadium more than doubled the price per square meter, attracted developers, led to thousands of residential units, and changed the neighborhood, but mobility barriers still hinder the area around the arena.
-
Brazil delivers a giant dam worth R$ 365.7 million in Rio Grande do Sul, capable of forming an artificial lake with 138 million m³ of water, equivalent to 55,300 Olympic swimming pools and an area larger than 2,500 football fields.
-
Global tsunami of risk aversion may hit the national economy hard. With negative macroeconomic fundamentals (inconsistent fiscal trajectory, prohibitive credit costs, uncontrolled inflation, currency appreciation, and flight of direct investment), the country may be left out of foreign investment opportunities.
-
While millions receive Bolsa Família and work in the informal sector, one fact puzzles experts: why do even affordable contributions to the INSS fail to convince Brazilians to plan for retirement?
But the case shows that rapid adoption can also generate a bill that is difficult to sustain. The company has not abandoned AI, but decided to implement governance over the use of tools. The stated goal is to maintain scale adoption without allowing costs to advance uncontrollably.
Annual budget ended in four months
According to the report, Uber acknowledged that the funds allocated for artificial intelligence in 2026 had already been fully used even before the end of the first half of the year. In April, the company’s Chief Technology Officer, Praveen Neppalli Naga, stated that the annual budget allocated to the area was exhausted.
This data explains the change in stance. If AI promises to accelerate tasks, write code, support decisions, and reduce work steps, it also demands high computational power. In practice, each query, agent, or tool used at scale can become a recurring expense.
Developers are the most affected by the rule
The new limit mainly affects AI-based programming software. Tools like Claude Code and Cursor can help developers write, review, and speed up code snippets, but intensive use increases the bill.
Each employee now has a monthly cap of $1,500 per tool. If they need to spend more, they will have to justify the use and obtain approval. This type of control indicates that corporate AI has entered a less exciting but more decisive stage: the accountability phase.
Artificial intelligence remains strategic for the company
Despite the spending halt, Uber states that the measure does not represent a withdrawal from the technology. On the contrary, the company continues to treat artificial intelligence as an important part of its operation and seeks a considered responsible and sustainable adoption.
Company executives have highlighted, in recent months, the growth in the use of these solutions in various areas. CEO Dara Khosrowshahi even stated that about 10% of the company’s code was already being produced with the help of AI agents.
Productivity still needs to become a measurable result
The great dilemma is proving whether the promised productivity really pays off. Artificial intelligence tools can accelerate tasks, reduce development time, and help teams solve problems more quickly.
But turning this gain into concrete financial results is not simple. In May, Uber’s Chief Operating Officer, Andrew Macdonald, acknowledged that it is still difficult to establish a direct relationship between increased AI use and the delivery of more products or features to customers.
Companies discover that AI also requires governance
The Uber episode reveals a larger movement in the market. After a phase marked by enthusiasm and accelerated adoption, companies are beginning to create mechanisms to measure usage, control costs, and define who can spend more on artificial intelligence tools.
This control does not mean rejection of the technology. It means that AI is leaving the experimental field and entering the real budget of companies, where each tool needs to compete for priority with other areas, goals, and investments.
Computational cost weighs on the final account
Advanced models require great processing capacity. This makes intensive use more expensive, especially when thousands of employees start using tools simultaneously in daily tasks.
The problem is that the cost does not appear only in the initial hiring. It grows as usage increases. The more AI integrates into daily work, the more important it becomes to understand who uses it, how much they use it, and what return this use delivers.
Limit per employee changes internal culture
By creating a monthly cap per employee, Uber turns the use of AI into a more conscious decision. Developers and teams need to evaluate when the tool really adds value and when it is just being used for convenience.
This type of rule can change internal behavior. From the moment there is a limit and approval for extra expenses, artificial intelligence ceases to be treated as an unlimited resource and starts being seen as expensive infrastructure that requires criteria.
The challenge is to measure return without stifling innovation
Controlling costs too much can reduce experimentation. Allowing unlimited spending can blow budgets. This is the difficult balance that companies like Uber are trying to find in AI adoption.
The central issue is not just cutting expenses, but understanding which uses really generate productivity, speed, and business impact. The next stage of the race for artificial intelligence may be less about who adopts first and more about who can prove return.
Uber case becomes a warning for other companies
A Uber is not the only company reviewing AI expenses. According to the report, other companies have also started adopting control mechanisms as these tools become part of the corporate routine.
This indicates that the market is entering a more mature phase. The initial euphoria still exists, but now it coexists with spreadsheets, limits, approvals, and objective questions about cost-benefit.
The technology continues to advance, but the bill also grows. Do you think companies should allow the use of AI to accelerate teams or create strict limits before expenses get out of control? Leave your opinion in the comments.
Be the first to react!