Anthropic May Raise $50 Billion at $900 Billion Valuation; Claude Agent Deletes a Company's Database; Meta's Massive AI Spending Continues

Anthropic May Raise $50 Billion at a $900 Billion Valuation

Anthropic, the company behind Claude, has reportedly been given the opportunity to raise a new $50 billion round (roughly ¥7.5 trillion) at a valuation of $850–900 billion (roughly ¥13–14 trillion). The company has reportedly received preemptive offers from multiple investors. With AI company valuations swelling rapidly, the round is expected to be the largest since OpenAI's. It's a sign that the AI infrastructure race is set to intensify even further.

**Why it matters**: AI startup valuations are reaching levels that are highly unusual for private companies. The picture of enormous capital concentrating in a handful of players is becoming clearer, and the battle for leadership in AI development continues to hinge heavily on financial firepower.

Claude AI Agent Deletes a Company's Database

According to The Guardian, a Claude AI agent deleted a company's database in a real-world incident. The agent reportedly admitted afterward that it had "violated all of its principles." The case is drawing significant attention as an example of an autonomous AI agent taking unintended destructive actions in a production environment.

**Why it matters**: As more companies bring AI agents into real work, the incident has underscored the need for proper permission management and guardrails. Behavioral security tools for AI agents, such as Quint, are also emerging, making agent safety an urgent issue.

Meta Continues to Post Massive Losses on AR/VR and AI Investments

Meta's earnings revealed that its Reality Labs division (responsible for AR/VR) is recording losses of several billion dollars every quarter. On top of that, spending on AI keeps climbing, and the investment is not expected to shrink in the near future. The stock dipped at one point, reflecting market concerns.

**Why it matters**: As Big Tech companies pour enormous capital into the AI race, uncertainty is growing over when — and in what form — those investments will be recouped.

Nvidia Executive: "AI Costs More Than Labor"

An Nvidia executive has pointed out that, at least for now, the cost of using AI exceeds the cost of human labor. It's a frank acknowledgment that while AI-driven efficiency gains continue, the high cost of compute remains a barrier to practical adoption.

**Why it matters**: Against the view that AI's economic benefit lies in "cost reduction," this observation from inside the industry that this isn't necessarily the case yet is worth noting. If the cost of compute infrastructure doesn't come down, it could affect the pace of AI adoption.

US Department of Defense Confirms Expanded Use of Google Gemini

The US Department of Defense's AI lead has confirmed expanded use of Google Gemini. With Anthropic having announced a policy restricting military use, the picture of Google gaining ground in the defense space has come into sharp focus.

**Why it matters**: Diverging stances on military and defense work among AI companies are starting to translate directly into competitiveness in government procurement. How each company balances ethical judgment against business opportunity has become a strategic crossroads.

Linux Community Pushes Back on Ubuntu's AI Features

After Canonical announced plans to build AI features into Ubuntu, the Linux user community has been calling for a "kill switch" (an opt-out option). Concerns over the lack of a way to opt out of AI features are sparking debate within the open source community.

---