Anthropic becomes the world's most valuable AI startup, and files to go public

Good morning, builders 👋 Anthropic just leapfrogged OpenAI to become the most valuable AI startup on the planet, and then quietly filed to go public. The frontier-lab pecking order just got rewritten.

Today:

• Anthropic's $965B moment (and its surprise IPO filing)
• A token-billing shake-up that's spooking every Copilot power user
• 5 tools worth a look + a prompt for pressure-testing your roadmap

In today's Inovaite:

• ⚡ Anthropic hits $965B, files for IPO, passes OpenAI
• 🧰 How to read a confidential S-1 like an investor
• 🛠️ 5 trending AI tools
• 📰 6 quick hits
• 💬 A prompt worth stealing

⚡ THE LEAD

💰 Anthropic becomes the world's most valuable AI startup, and files to go public

What happened: Anthropic closed a $65B Series H at a $965B post-money valuation (Anthropic), surpassing OpenAI as the most valuable private AI company. Then, days later, it confidentially filed a draft S-1 with the SEC (Fortune), setting up a possible public listing as soon as fall 2026.

Why it matters: This is the largest AI funding event of the cycle, and the first time Anthropic has eclipsed OpenAI's ~$852B private mark (Al Jazeera). Its valuation roughly tripled from ~$380B in February. The near-simultaneous IPO move signals the AI capital race is shifting from private megarounds to public markets, where quarterly scrutiny will reshape frontier-lab economics. Lead investors include Altimeter, Dragoneer, Greenoaks, and Sequoia (Crunchbase).

The takeaway: If you build on Claude, your core vendor is about to become a public company. Expect more enterprise muscle, more compute, and more pressure to monetize. Plan accordingly.

🧰 BUILD WITH IT

Read a confidential S-1 like an investor (in 10 minutes)

1. Skip the boilerplate. Jump straight to "Risk Factors" and "MD&A" (Management Discussion & Analysis); that's where the real story lives.

2. Find the revenue recognition and cost of revenue notes; for an AI lab, compute spend hides here.

3. Check customer concentration disclosures: how dependent is the business on a few big accounts?

4. Drop the doc into an AI assistant and run the prompt below to surface what the narrative is downplaying.

Prompt: You are a skeptical institutional analyst. Here is an S-1 excerpt. List the 5 things management is most likely downplaying, each with the specific sentence that tipped you off.

🛠️ TOOLS: 5 trending right now

• 🤝 Mina Meeting Assistant: your AI teammate now responds and executes actions live during your calls. link
• 📊 Databox MCP: chat with your business data inside Claude, ChatGPT, and more. link
• 🔎 Tabstack Web Research: run a research agent that returns cited answers in a single API call. link
• 🛠️ Mistral Vibe: Mistral's agent for long-running, multi-step work and coding. link
• 💸 Tokenwise: a smart LLM proxy that shows where you're overpaying on AI spend. link

📰 QUICK HITS

• Cognition raised $1B at a $26B post-money valuation for Devin, its autonomous software engineer. Co-led by Lux Capital, General Catalyst, and 8VC, on a reported ~$492M revenue run-rate. link
• SoftBank committed up to €75B (~$87B) to build 5 GW of AI data-center capacity in France, its biggest European infrastructure bet yet. link
• NVIDIA unveiled Nemotron 3 Ultra at Computex: a 550B-param (55B active) open-weights model it's billing as the most capable US open model; weights land on Hugging Face this week. link
• GitHub Copilot moved to usage/token-based billing for all plans on June 1, sparking developer backlash as power users project steep cost jumps on agentic sessions. link
• Microsoft & Google are moving to challenge Anthropic and OpenAI with their own AI coding models. Microsoft is expected to unveil a cost-focused Copilot coding model at Build 2026 (keynote June 2). link
• ServiceNow + Accenture launched a Forward Deployed Engineering program to push enterprise agentic AI from pilot to production. link

💬 PROMPT

Act as a contrarian VC. Given this AI product idea: [paste].
1. Name the 3 strongest reasons it wins.
2. Name the 3 most likely reasons it dies in 18 months.
3. Tell me the single experiment that would most cheaply prove or kill it this month.
Be blunt. No hedging.

That's a wrap. Build on. — Inovaite

How was today's issue? 😍 / 🙂 / 😴