Uber’s Anthropic AI push hits a wall
anthropic claude
| Source: HN | Original article
Uber’s internal push to embed Anthropic’s AI tools has run out of steam. Chief Technology Officer Praveen Neppalli Naga told The Information that the ride‑hailing giant has already exhausted its 2026 AI budget – a $3.4 billion R&D allocation – within the first quarter of the year. The shortfall stems from a surge in the use of Anthropic’s Claude Code, a generative‑coding assistant that teams have adopted for everything from route‑optimization scripts to fraud‑detection pipelines.
The overspend forces Uber back to the drawing board, with the company now reassessing how it scales AI‑driven features without overrunning costs. As we reported on April 19, Anthropic’s Claude Code was recently exposed in a leak that highlighted critical command‑injection vulnerabilities. Those security concerns, combined with the tool’s high per‑token pricing, appear to have amplified Uber’s fiscal strain.
Why it matters goes beyond a single corporate budget. Uber’s experience underscores a growing industry tension: the promise of rapid AI‑enabled innovation versus the reality of steep, often unpredictable, operating expenses. For firms that have bet heavily on third‑party large‑language models, the episode serves as a cautionary tale about hidden consumption spikes and the need for tighter cost‑control mechanisms. It also puts pressure on Anthropic, whose pricing model may now face scrutiny from other enterprise customers wary of runaway spend.
What to watch next is whether Uber renegotiates its contract with Anthropic, pivots to an in‑house model, or throttles AI deployment across its product stack. Anthropic’s response—potentially adjusting pricing tiers or offering more granular usage analytics—will be a key indicator of how the market adapts to enterprise cost concerns. Finally, other AI‑heavy players such as Lyft, DoorDash and Amazon are likely to monitor Uber’s recalibration closely, as they chart their own paths through the same budgetary minefield.
Sources
Back to AIPULSEN