Anthropic prices its IPO at $83 a share
Anthropic priced its stock at $83 a share today, which gives every founder a public price tag on a frontier lab. The headline matters less than the filings that come after it. Its compute spending and pricing will now be read line by line by anyone who sells to or competes with Claude. Use that benchmark when you think about your own valuation and pricing.
- Anthropic prices its IPO at $83 a share
- Gemini 4 Neon ranks third, with a middle price
- Dippr lets coding agents build internal tools
- Nvidia halts H200 shipments to China
- Five quick hits and today's AI Index move
Top stories
All news →What happened: Anthropic priced its Class A shares at $83 each, above its initial range, ahead of a Nasdaq listing.
- The offering covers 980,000,000 shares in total: 956,351,354 from the company and 23,648,646 from selling stockholders, so Anthropic gets no money from the second group.
- Underwriters (the banks that sell the shares) hold a 30-day option to buy up to 147,000,000 more.
- Trading will run on the Nasdaq Global Select Market, with the offering expected to close on November 20.
Why it matters: A public listing gives founders a valuation benchmark for a frontier lab. Its compute spending and risk disclosures will now appear in quarterly filings, so watch how it talks about pricing and enterprise sales.
What happened: Google released Gemini 4 Neon on November 18, a model that sits between Carbon for long, demanding projects and Hydrogen for fast everyday work.
- Neon ranks third on the Artificial Analysis Intelligence Index with a score of 61 at its highest thinking setting, and second on the Vals Index. It offers four thinking levels and a 1 million-token context window (how much text it can read at once).
- Developers can reach it through the Gemini API and Google AI Studio. AI Pro and Ultra subscribers can pick it in the Gemini app, where it replaces Argon in the model picker and Gemini 3.1 Pro is retired.
- List price is $3 per million input tokens and $15 per million output tokens, with cached input at $0.15. Carbon and Argon are discounted to $2 and $10 until November 30, then rise to $4 and $20.
Why it matters: If you build on Gemini, test whether Neon can handle work you now send to a pricier model. Carbon and Argon return to higher prices after November 30, so your cost picture will shift then.
What happened: Dippr, listed on Product Hunt on November 18, lets AI coding agents build and run a company's internal software using its own data.
- Setup is one step: paste a prompt into a coding agent such as Claude Code, Codex or Cursor, and it installs a command-line tool called dip plus Dippr's MCP server (a connector that lets AI tools reach outside data).
- Each data source connects once. The site names Snowflake, Salesforce, Zendesk, Postgres and Stripe as options, and metric definitions are set once so every agent calculates them the same way.
- The Product Hunt listing shows 207 upvotes and 38 comments.
Why it matters: This points to where internal tools are heading: agents build them on top of company data. If you sell internal software or data tooling, you need a clear reason a customer should pay you instead of prompting an agent.
Quick hits
- Nvidia has stopped shipping licensed H200 chips to China, and its next-quarter outlook assumes none. Reuters via TimesLIVE →
- AI agents out-questioned KPMG accountants in an audit trial, according to europeanbusinessmagazine.com. europeanbusinessmagazine.com →
- Microsoft AI just released seven new models. microsoft.ai →
- Toone lets you build AI workflows by describing them in plain language, using ChatGPT or Claude. Product Hunt →
- Wherobots launched Waldo, an AI agent that answers questions about maps and satellite imagery. Product Hunt →


