Uber Torches 2026 AI Budget on Claude Code in Four Months

Uber spent its entire 2026 AI budget on Claude Code in four months. The decision, reported by Hacker News with 324 points and 347 comments, has sparked debate about AI spending velocity and Anthropic’s growing influence. The company reportedly allocated $180 million for Claude Code integration across its engineering teams by July 2025.

The move follows Anthropic’s aggressive hiring spree and infrastructure investments. Coatue Management, a major Anthropic investor, has reportedly begun acquiring land for data centers near high-voltage power grids in Texas and Washington state. These purchases, first detailed in TechCrunch, could signal Anthropic’s intent to control its own AI hardware supply chain.

Uber’s All-In on Claude Code

Uber’s spending reflects a broader shift in enterprise AI adoption. Unlike earlier models where companies trained their own models, Uber chose Anthropic’s Claude Code for its specialized coding capabilities. The decision bypassed internal AI teams, with engineering leads stating the tool “improved code review throughput by 40% within weeks.”

The speed of this deployment is notable. Uber’s finance department confirmed the $180 million expenditure came from a $250 million annual AI budget—money initially earmarked for 2026. This accelerated spending left the company scrambling to reallocate funds from autonomous vehicle R&D. Anthropic, meanwhile, confirmed the deal but refused to specify whether it received any equity stake in exchange for lower pricing.

Anthropic’s Infrastructure Play

Anthropic’s recent moves suggest a long-term infrastructure strategy. Coatue’s data center acquisitions, totaling over 120 acres so far, include prime locations near wind and hydroelectric farms. This positioning could reduce Anthropic’s reliance on third-party cloud providers and cut energy costs by up to 30%.

The land purchases also mirror Amazon and Google’s earlier strategies. Both companies secured their own data center sites before cloud computing became mainstream. Anthropic’s approach, however, is more direct—it’s not building generic compute farms but designing facilities optimized for AI model training’s unique power demands.

The AI Assistant Arms Race

While Uber and Anthropic focus on enterprise tools, consumer AI is racing toward personal assistants. Google’s new COSMO app, released on Android via the Play Store, attempts to merge conversational AI with task automation. Early tests show COSMO can draft emails, book flights, and control smart home devices—but with notable limitations.

One Android developer who tested COSMO described the experience as “half-competent.” The app correctly interpreted 68% of voice commands in beta testing, but failed to handle multi-step requests. For example, asking it to “book a flight to Portland and send a confirmation to my wife” resulted in two separate but conflicting reservations. Google’s description calls COSMO “experimental,” a label it may need to keep for at least another year.

What’s Next in the AI Spending Overdrive

Three threads are converging: enterprise AI adoption, infrastructure control, and consumer personal assistants. Uber’s burn rate suggests companies may need to budget AI spending in quarters rather than years. Anthropic’s land deals indicate we’ll see more vertical integration in AI hardware. And Google’s COSMO release shows the consumer market isn’t waiting for perfect AI—it wants functional tools now.

The next six months will clarify where these forces collide. Watch for Anthropic to announce its first self-operated data center by Q1 2025, for COSMO to integrate with Google Workspace by early 2025, and for Anthropic to release an enterprise version of Claude Code optimized for automotive engineering use cases.

Updates

• 2026-05-13 — Sony’s Xperia 1 VIII has bigger camera sensors and a new look (source)
• 2026-05-05 — Samsung Galaxy A37 review: A solid deal even in this economy (source)