Presented by
Good morning, AI enthusiasts. Anthropic just sued the US Department of Defense — and within hours, its biggest competitors' employees were rushing to its defense.
If the government can punish a US firm for disagreeing with its safety stance, the precedent reaches far beyond one company. Has Anthropic just drawn a line that changes the whole industry?
In today's recap:
Microsoft's Copilot Cowork executes M365 tasks for you
ChatGPT hits 900M users, Claude growing 200% annually
Anthropic sues the Pentagon over supply-chain label
3 new AI tools, prompts, and more
ANTHROPIC
Anthropic takes the Pentagon to federal court
Getty / Futurism
Recaply: Anthropic just filed two federal lawsuits against the Defense Department after the Pentagon labeled it a supply-chain risk, calling the move "unprecedented and unlawful" and arguing it's being punished for its protected speech.
Key details:
Anthropic filed one complaint in San Francisco federal court and a second in the DC Circuit Court of Appeals, seeking to invalidate the supply-chain designation and pause enforcement while the case proceeds.
Anthropic named the Department of War and more than a dozen other federal agencies as defendants, with the company estimating the actions could cost it hundreds of millions of dollars in the near-term.
More than 30 OpenAI and Google DeepMind employees, including Chief Scientist Jeff Dean, filed an amicus brief in support, calling the designation "an improper and arbitrary use of power."
Both complaints were filed Monday, weeks after Defense Secretary Hegseth ordered federal agencies to stop using Claude on Feb. 27, with a six-month transition window given to shift to other providers.
Why it matters: There's something remarkable happening here, and it's not just the lawsuit. OpenAI and Google employees rushed to file legal support for Anthropic within hours of the complaints dropping. That doesn't happen in normal market competition. The precedent this case sets matters for every AI company. If the government can punish a US firm for its safety stance, the chilling effect could reach far beyond Anthropic.
PRESENTED BY MINTLIFY
Ship Docs Your Team Is Actually Proud Of
Mintlify helps you create fast, beautiful docs that developers actually enjoy using. Write in markdown, sync with your repo, and deploy in minutes. Built-in components handle search, navigation, API references, and interactive examples out of the box, so you can focus on clear content instead of custom infrastructure.
Automatic versioning, analytics, and AI powered search make it easy to scale as your product grows. Your docs stay accurate automatically with AI-powered workflows with every pull request.
Whether you're a dev, technical writer, part of devrel, and beyond, Mintlify fits into the way you already work and helps your documentation keep pace with your product.
MICROSOFT
Microsoft's Copilot moves from chat to action
Microsoft / Screenshot
Recaply: Microsoft just launched Copilot Cowork, a cloud AI agent that plans and executes multi-step tasks across M365, built with the technology behind Anthropic's Claude Cowork.
Key details:
Users describe an outcome and Cowork breaks it into a plan, running tasks in the background across Outlook, Teams, Excel, and other M365 apps, with checkpoints for user approval.
Microsoft 365 E7, a new bundle including Copilot, Agent 365, and full security, launches May 1 at $99 per user per month; Anthropic's Claude Cowork launch triggered a $285B selloff in enterprise software stocks earlier this year.
Microsoft is also making Claude models available directly in Copilot Chat for Frontier users, positioning Anthropic's AI as a general-purpose option alongside OpenAI's models, according to VentureBeat.
Copilot Cowork is in limited Research Preview now, with broader access coming through the Frontier program in late March 2026.
Why it matters: Microsoft's move to make Copilot an execution engine, not just a chat assistant, is one of the shrewder plays in enterprise AI right now. With Anthropic's Claude technology running inside M365's security perimeter, they don't have to choose between intelligence and governance. The real question is whether users actually trust an AI to act on their behalf. That's a behavioral shift no product launch can shortcut.
AI RESEARCH
ChatGPT now reaches one in ten people globally
a16z
Recaply: a16z just published its 6th Top 100 Gen AI Consumer Apps report, finding ChatGPT reaches 900 million weekly users while Claude and Gemini are growing paid subscribers at over 200% annually.
Key details:
Rankings track monthly unique web visits and mobile MAU, with this edition broadening its scope to include AI-powered apps like Canva, CapCut, and Notion alongside AI-native products.
ChatGPT is 2.7x larger than Gemini on web and 2.5x on mobile, adding 500 million weekly active users in a single year; over 10% of the global population now uses it weekly.
Singapore ranks first in per capita AI adoption globally; the US, which produced most of these products, ranks 20th, according to a16z's index combining web visits and mobile MAU per capita.
Rankings are based on January 2026 data from SimilarWeb and Sensor Tower; published March 9, 2026 as the 6th edition of this bi-annual series.
Why it matters: There's been talk of an AI race that's already decided. But Claude grew paid subscribers over 200% last year, and Gemini grew 258%. ChatGPT is still 8x larger than Claude on this metric, but the trajectory tells a different story. Context compounds: the more a model knows about you, the harder it gets to switch. Who leads today matters less than who you're locked into tomorrow.
Also worth noting: Claude Code hit $1B in annualized revenue in just six months. AI is quietly moving beyond browsers and into the OS itself. Developer tools, desktop agents, and notetakers now have a combined audience that doesn't show up in traditional web traffic data. The real usage numbers for AI may be significantly larger than any top-100 list can capture.
NEWS
What Matters in AI Right Now?
Anthropic introduced Code Review in Claude Code, a multi-agent system that flags logic errors in pull requests for Teams and Enterprise customers, with early data showing substantive reviews jumping from 16% to 54% at $15-25 per review.
OpenAI acquired Promptfoo, an AI red-teaming and security platform used by 25%+ of Fortune 500 companies, with its open-source project staying free and the technology being integrated into OpenAI's Frontier platform.
Cursor launched Automations, always-on background coding agents that run on schedules or event triggers inside a cloud sandbox, with native integrations for Slack, GitHub, Linear, and PagerDuty.
Luma launched Luma Agents, a platform that handles end-to-end creative work across text, images, video, and audio using its new Uni-1 multimodal model, with customers including Publicis Groupe, Adidas, and Mazda already using it via API.
Lightricks released LTX-2.3, a major upgrade to its open-source video model with sharper detail, stronger motion coherence, and native vertical format support, alongside LTX Desktop, a free local video editor built directly on the same engine.
Tencent and Zhipu launched OpenClaw-powered AI agents, with Tencent's WorkBuddy and Zhipu's AutoClaw driving stock gains of 6.2% and 16% respectively, as Chinese tech firms race to release local versions of the viral agentic framework.
Yann LeCun's AMI raised $1.03B in seed funding at a $3.5B pre-money valuation, co-led by Cathay Innovation, Greycroft, and Bezos Expeditions, to build world models as an alternative to large language models.
OpenAI partnered with McKinsey, BCG, Accenture, and Capgemini to push its Frontier AI agent platform, with rival Anthropic also signing deals with Deloitte and Accenture as both companies expand their enterprise reach through consulting firms.
TOOLS
Trending AI Tools
⚙️ Code Review - Anthropic's multi-agent PR reviewer
⚡ Copilot Cowork - Microsoft's OpenClaw-powered AI agents
🔍 Perplexity Computer - Perplexity's OpenClaw-powered AI agents
EVENTS
Voice AI HackSprint 2.0: March 14, 2026 • San Francisco, CA
NVIDIA GTC 2026: March 16-19, 2026 • San Jose, CA
ClawCon: March, April, and May, 2026
🧡 Enjoyed this issue?
🤝 Recommend our newsletter or leave a feedback.
How'd you like today's newsletter?
Cheers, Jason
