Anthropic draws lines with the Pentagon. Meta moves its chip bets. Nvidia prints another monster quarter. China targets critics with AI tools and tries to siphon model capability. Same theme: AI advantage is being fought on access, compute, and control.

WHAT YOU’LL LEARN

- Why Anthropic’s “no” to unfettered Pentagon access is a major precedent for defense AI contracting

- What the Meta–Google multibillion AI chip partnership signals about TPUs versus GPUs

- What Nvidia’s quarter suggests about demand durability, and where the China revenue picture tightens

- How China is using ChatGPT to target Japan’s PM and stealing info from Claude

- What “rogue agent” research is trying to solve, and why permissions are the real danger zone

- Why circular AI funding can amplify systemic risk if revenue assumptions wobble

TIMESTAMPS (CHAPTERS)00:00 – AI chip wars, Anthropic vs Pentagon, and China advantage plays

00:16 – Anthropic says no: unrestricted Claude access dispute

00:39 – The red lines: mass surveillance and autonomous weapons

01:05 – Pentagon options: Defense Production Act vs blacklisting

01:53 – Meta and Google strike a multibillion AI chip deal (TPUs)

02:24 – Meta hedges across Google, AMD, and Nvidia; internal chip effort struggles

02:51 – Nvidia crushes forecasts: $68B Q4 sales and demand dynamics

03:24 – China angle: Nvidia revenue in China and the broader chip contest

04:11 – The strategic debate: dependence theory vs China domestic chip scale

05:15 – OpenAI report: China-linked misuse of ChatGPT for targeting and influence

06:15 – Anthropic vs China: alleged Claude distillation attempts by AI labs

07:37 – Why distillation becomes a national security issue in practice

08:03 – Anthropic research: rogue agents, hijacking risk, and model steering

09:02 – OpenAI funding round and the circular-capital question

09:34 – Closing

SOURCES

- Anthropic Rejects Pentagon AI Work — Ethics Over Defense: https://apnews.com/article/anthropic-pentagon-ai-hegseth-dario-amodei-b72d1894bc842d9acf026df3867bee8a#

- Trump Bans U.S. Government From Using Anthropic AI After Pentagon Clash: https://www.cnbc.com/2026/02/27/trump-anthropic-ai-pentagon.html

- NVIDIA’s Record Profits — Agentic AI Drives Historic $68B Quarter: https://finance.yahoo.com/news/ai-bubble-nvidia-posts-record-224949246.html

- NVIDIA Revenue Soars — Has Agentic AI Hit a Turning Point?: https://aimagazine.com/news/nvidia-revenue-soars-has-agentic-ai-hit-a-turning-point

- Meta Scraps In-House AI Training Chip After Design Struggles: https://thein.fo/4chhFPC

- Meta Turns to Google TPUs as Advanced AI Chip Effort Stalls: https://biz.chosun.com/en/en-it/2026/02/27/4M67TQTJ25BGXKM7VR6X24KNXM/

- AI Threat Report — How Bad Actors Exploit OpenAI Tools: https://cdn.openai.com/pdf/df438d70-e3fe-4a6c-a403-ff632def8f79/disrupting-malicious-uses-of-ai.pdf

- Anthropic Accuses Chinese Labs of Stealing AI via Model Distillation: https://cyberscoop.com/anthropic-accuses-chinese-labs-ai-distillation-cyber-risk/?utm_campaign=CyberScoop%20-%20Editorial&utm_content=371245330&utm_medium=social&utm_source=linkedin&hss_channel=lcp-11061843

- Elon Musk Claims Anthropic Stole AI Training Data at Massive Scale: https://finance.yahoo.com/news/elon-musk-calls-anthropic-guilty-233018277.html

FOLLOW GOOD REVENUE

• YouTube: https://www.youtube.com/channel/UCHrhuWbpYnUwrR75H2Ip4yQ 
• Spotify: https://open.spotify.com/show/44me6PABSQ0k0GfphM70au
• Apple Podcast: https://podcasts.apple.com/us/podcast/good-revenue/id1707461473 

Hosted on Acast. See acast.com/privacy for more information.