Urgent.News

What's breaking now, across thousands of outlets.

AI

Anthropic flags billions in losses and AI risks ahead of IPO as Dario Amodei urges caution

Anthropic's IPO prospectus reveals a net loss of $42 billion in 2025, highlighting the high costs of AI scaling. The company reported a 12-fold revenue increase to nearly $4.6 billion, despite operating losses exceeding $8 billion. CEO Dario Amodei emphasizes the need for caution in AI deployment while addressing competitive pressures. Anthropic aims for a valuation exceeding $2 trillion in its…

Anthropic flags billions in losses and AI risks ahead of IPO as Dario Amodei urges caution

Anthropic, the AI giant, has disclosed staggering financial losses and significant AI risks in its IPO prospectus. In 2025, the company reported a net loss of $42 billion, including a $34 billion accounting charge, while operating losses exceeded $8 billion. Revenue surged 12-fold to nearly $4.6 billion. The company spent $7.33 billion on compute and infrastructure, a threefold increase from the previous year, accounting for more than half of its $12.65 billion operating expenses.

Looking ahead, Anthropic has committed to $518 billion in cloud and infrastructure obligations. CEO Dario Amodei has expressed concerns about the potential risks of increasingly autonomous AI models, which can behave in unexpected and potentially harmful ways. Despite these risks, Anthropic is targeting a public valuation of more than $2 trillion, more than double its private valuation of $965 billion.

The company aims to list after the US midterm elections, which is likely to bring public market investors into an AI investment race.

Brief written by urgent.news from Times of India's own syndicated text. Machine-written — may contain errors; check the original before relying on it.

Read the original at timesofindia.indiatimes.com →

More in AI

"Is There Any Point to This?" Removing the Jev Plugins I Added to Claude Code After One Week

At the end of my previous article , I wrote what I would do next. I would run the plugin that suggests skills with Jev in a setting where it adds nothing and only writes a log, and match the skills…

  • Jev plugin logged 1,242 skill selections over one week.
  • Only 28 instances of Jev's suggestions were executed within 30 minutes.
  • Removing Jev plugin halted Claude Code 8 times due to incomplete task verification.

🧠 Building Living City: An AI-Powered City Intelligence Platform with Hindsight Memory

# 🌆 Living City: What If a City Could Remember? Cities are constantly changing. Weather changes. Air quality changes. Events appear and disappear. People move through the city.

  • Living City is an AI-powered city intelligence platform with Hindsight memory.
  • Hindsight is a persistent memory layer that gives cities a memory.
  • Living City's core loop connects past experiences to current decisions.

More from Tuesday 29 September →