Uber Exceeds Its AI Budget, Meta Exploits Internal Data

Uber Exceeds Its AI Budget Well Before 2026

According to The Information, Uber Technologies has already exhausted its entire budget allocated for artificial intelligence for the year 2026, just a few months into the fiscal year. Praveen Neppalli Naga, the company's Chief Technology Officer, explained that the intensive use of code generation tools, particularly Claude Code developed by Anthropic, has accelerated this situation.

In a recent interview, Praveen Neppalli Naga described an organization where AI is deployed at all levels. This integration is evident through the acceleration of software development, the automation of customer service via agents, and direct integration into products intended for drivers and users. This "AI-first" approach transforms AI into a continuously consumed resource rather than a mere one-time investment.

The heavy reliance on code generation tools has become essential for Uber, especially for fixing bugs and iterating quickly. However, this dependence complicates cost management, as the more AI enhances productivity, the more it increases resource consumption. In a company the size of Uber, this rebound effect is particularly challenging to control.

For Uber, economic optimization takes a backseat to competitive advantage. The exceeding of the AI budget is thus seen as a strategic choice, prioritizing technological acceleration over cost control. A fundamental question remains: Can Uber maintain this intensive usage model without fundamentally rethinking its stack, or even without internalizing some of these capabilities to regain economic control?

Meta Uses Internal Data for Its AIs

Meta Platforms has decided to collect mouse movements, clicks, and keystrokes from its employees on certain internal tools to train its artificial intelligence models, according to Reuters. This initiative aims to develop agents capable of replicating real user behaviors, based on sequences of actions rather than content.

This approach marks a transition towards execution AI, where behavioral data becomes central. Meta assures that this data will not be used to evaluate employees individually but does not rule out the possibility that it could be used to automate certain tasks currently performed by employees.

At the same time, Meta plans to cut 8,000 jobs starting May 20, representing nearly 10% of its workforce, as part of a reorganization focused on artificial intelligence.

Emmanuel Marill Appointed Head of OpenAI for the EMEA Region

OpenAI has appointed Emmanuel Marill as the General Manager for the EMEA region, a newly created role based in Paris. A former executive at Airbnb, Marill will be responsible for promoting the adoption of ChatGPT in a region that already boasts over 120 million monthly users.

With experience at LVMH, Roland Berger, Groupon, and Meta Platforms, Marill will need to navigate a complex European environment marked by the Digital Services Act and the AI Act. He will also have to strengthen OpenAI's position against competitors like Anthropic.