The Pentagon's standoff with Anthropic over military access to Claude dominates today's headlines, while a Trump administration official claims China's DeepSeek trained its latest model on banned Nvidia Blackwell chips. India's AI Impact Summit wraps up with 88 nations signing a cooperation pact, and regulators across multiple jurisdictions push new rules on AI-generated content, child safety, and platform accountability.

Pentagon Summons Anthropic CEO Over Military Use of Claude

Defense Secretary Pete Hegseth has called Anthropic CEO Dario Amodei to the Pentagon for what sources describe as a tense meeting over the military's access to Claude. Hegseth has reportedly threatened to designate Anthropic a "supply chain risk" if the company continues to restrict how its AI can be used on classified systems.

Scoop: Hegseth to Meet Anthropic CEO as Pentagon Threatens Banishment — Axios
Axios reports that Defense Secretary Hegseth summoned Dario Amodei for Tuesday talks, threatening to label Anthropic a supply chain risk over restrictions on Claude's military deployment.

Defense Secretary Summons Anthropic's Amodei Over Military Use of Claude — TechCrunch
TechCrunch confirms the high-stakes meeting and notes that military personnel already rely on Claude through Palantir for classified data analysis, making a ban potentially counterproductive.

Why it matters: This confrontation tests whether AI companies can maintain ethical guardrails on military applications when facing direct pressure from the U.S. government.

DeepSeek Allegedly Trained on Banned Nvidia Blackwell Chips

A senior Trump administration official claims China's DeepSeek used Nvidia's most advanced Blackwell AI chip to train its upcoming model, set for release as soon as next week — a potential violation of U.S. export controls.

US Claims China's DeepSeek Used Banned Nvidia Blackwell Chips — Benzinga
U.S. officials believe DeepSeek's next AI model was trained on Nvidia's restricted Blackwell chips, potentially breaching export controls designed to slow China's AI progress.

China's DeepSeek Trains AI on U.S. Nvidia Chip Despite Export Ban — Modern Diplomacy
The report raises questions about the effectiveness of U.S. chip export restrictions and how Chinese firms may be circumventing controls through intermediaries.

Why it matters: If confirmed, this would be the most significant known breach of U.S. AI chip export controls, escalating tech tensions between Washington and Beijing.

India AI Impact Summit — 88 Nations Sign Cooperation Pact

India's four-day AI Impact Summit attracted 250,000 visitors and executives from OpenAI, Anthropic, Google DeepMind, Nvidia, and Microsoft. The event culminated in 88 nations signing an AI cooperation agreement, with major investment commitments including Blackstone's $600 million stake in Indian AI infrastructure firm Neysa.

Beyond the Hype: 88 Nations Sign India's AI Pact as New Delhi Challenges Western Tech Monopoly — Indian Express
India positions itself as a global AI hub with 88 nations signing a cooperation pact, attended by Sundar Pichai, Sam Altman, Dario Amodei, Demis Hassabis, and Emmanuel Macron.

Chaos, Confusion and $200 Billion Dreams: What I Saw at India's AI Summit — CNBC
On-the-ground reporting captures the scale and ambition of the summit, with OpenAI announcing a partnership with Tata and Google expanding Gemini across Indian research institutions.

Why it matters: The summit signals India's emergence as a serious AI power broker, leveraging its 100 million ChatGPT users and diplomatic reach to attract Western investment and set global norms.

AI Policy & Content Regulation

AI Image Tools Must Follow Privacy Rules, Watchdogs Say — The Register
Britain's ICO and dozens of international regulators issued a joint statement demanding AI image generators comply with data protection laws, particularly around nonconsensual depictions of identifiable individuals.

X Is Working on 'Made with AI' Labels — The Verge
Elon Musk's X platform is developing disclosure toggles for AI-generated content, likely in response to India ordering platforms to adopt provenance systems like C2PA.

White House Working Against Florida Attempt to Limit AI for Minors — Daily Signal
The White House is reportedly lobbying against Florida's AI Bill of Rights, which would mandate parental controls on AI for minors, preferring a weaker federal framework instead.

Online Harms Bill Needs Framework for Reporting Threats in AI Chats, Experts Say — The Globe and Mail
Canadian experts urge the federal online harms bill to require AI companies to report credible threats to police, after OpenAI flagged but did not report content from the Tumbler Ridge shooter.

Why it matters: A wave of AI content and safety regulation is building across multiple jurisdictions simultaneously, with governments struggling to balance innovation with protection — particularly for children.

AI Industry & Investment

Big Tech to Invest About $650 Billion in AI in 2026, Bridgewater Says — Economic Times
Alphabet, Amazon, Meta, and Microsoft are collectively expected to spend $650 billion scaling AI infrastructure this year, according to a Bridgewater Associates analysis.

As AI Threatens Search, Profound Raises $96 Million to Help Brands Stay Visible — Fortune
The 18-month-old startup hits unicorn status with a $1 billion valuation, betting that brands will need to optimise how AI systems talk about them as traditional search declines.

Google VP Warns That Two Types of AI Startups May Not Survive — TechCrunch
Google's startup chief says LLM wrappers and AI aggregators have their "check engine light on," warning that thin layers over existing models won't sustain viable businesses.

Claude Code Targets COBOL, Sending IBM Stock Down Over 10% — Tech Startups
Anthropic's Claude Code announced it can automate legacy COBOL modernisation, triggering IBM's worst stock day since 2000 as investors question the future of its mainframe business.

Why it matters: The IBM sell-off shows how quickly AI capabilities can threaten entrenched enterprise moats, with a single product announcement wiping $30 billion in market value.

U.S. Launches Peace Corps-Backed 'Tech Corps' to Help Export AI, Counter China — CNBC
Washington creates a new Tech Corps modelled on the Peace Corps to deploy AI solutions in agriculture, education, and healthcare across developing nations, countering China's digital influence.

Generative AI Analyzes Medical Data Faster Than Human Research Teams — ScienceDaily
UCSF and Wayne State researchers found that generative AI could build medical prediction models in minutes that took human teams months, with some AI results matching or exceeding expert performance.

Relevant Resources
Claude — Anthropic's Thoughtful AI Companion — Background on the AI at the centre of the Pentagon dispute
AI Chips & Accelerators — Understanding the hardware behind export control tensions
Claude Code — Agentic Coding Tools — The tool shaking IBM's mainframe dominance
Agentic AI — How autonomous AI agents are reshaping industries
Data Centres & AI Superclusters — Where the $650 billion is going

Today's Pulse: 13 stories tracked across 14 sources — Axios, TechCrunch, Benzinga, Modern Diplomacy, Indian Express, CNBC, The Register, The Verge, Daily Signal, The Globe and Mail, Economic Times, Fortune, Tech Startups, ScienceDaily