Get tomorrow's brief in your inbox
Today: Anthropic's revenue exploded to $11.5 billion in Q2 (up 14x year-over-year) as it prepares for a potential fall IPO. The company also detailed how its new text watermarking system will work under EU AI Act compliance. SpaceX officially closed its $60 billion Cursor acquisition, giving the AI coding tool access to the world's largest GPU fleet.
Anthropic's Revenue Surges to $11.5B, IPO Looms
Anthropic reported preliminary Q2 2026 revenue of over $11.5 billion, up from $787 million in Q2 2025 and $4.73 billion just last quarter. The company also posted positive adjusted operating income for the first time. CFO Krishna Rao is leading early investor meetings ahead of a potential fall IPO.
What this means for you: Anthropic staying profitable and going public signals that Claude isn't going anywhere. If you've built workflows around Claude (coding assistance, content generation, customer support automation), this is validation that the platform will have the funding runway to keep improving. An IPO also typically means more enterprise features, better uptime SLAs, and expanded API access as the company courts institutional investors.
The practical takeaway: if you've been evaluating Claude vs. competitors for business use, this financial stability makes it a safer long-term bet. Expect pricing to remain competitive as Anthropic fights OpenAI for enterprise contracts.
SpaceX Closes $60B Cursor Acquisition
The AI coding editor Cursor is now officially part of SpaceX (which already absorbed xAI). Cursor says it will gain "access to the largest fleet of GPUs in the world," which should translate to faster completions and more capable models running behind the scenes.
If you use Cursor for development, expect performance improvements as SpaceX's compute infrastructure gets integrated. The risk: Cursor's roadmap now answers to Elon Musk's priorities. If you're uncomfortable with that or worried about vendor lock-in, VS Code with Copilot or Windsurf remain alternatives. No pricing changes announced yet.
CORS Chat: Quick Local LLM Testing UI
Simon Willison built CORS Chat, a browser-based UI for testing OpenAI-compatible chat endpoints. Works with LM Studio (with --cors flag) and OpenRouter. Conversations persist locally and export as JSON. It even renders SVG images progressively as they stream in.
If you're running local models via LM Studio or Ollama and want a quick way to test them without setting up a full app, this is worth bookmarking. Free, open source, zero setup beyond opening the page.
From One-Liner Prompts to Full Workflows
Inc. highlights how skilled AI users have moved beyond single-sentence prompts to building entire workflows with chatbots. The gap between casual users and power users is widening fast.
Action step: if you're still typing "write me an email about X," you're leaving value on the table. Structure your prompts with context, constraints, format requirements, and examples. Better yet, build reusable prompt templates for your recurring tasks (client proposals, weekly reports, social media calendars). The difference between a 30-second prompt and a 2-minute structured prompt is often a 10x improvement in output quality.
Claude Watermarking: What You Need to Know
Anthropic detailed how its upcoming text watermarking (required by the EU AI Act) will work. Key facts:
If you use Claude to draft client-facing content, this matters. Light editing (fixing a sentence here and there) won't strip the watermark. If you want fully unwatermarked text, you'll need to substantially rewrite Claude's output, which arguably makes it your writing anyway. For code generation, the impact is negligible.
AI Safety Under Scrutiny as Grok Abuse Case Escalates
A woman joined a lawsuit against xAI alleging her stepfather used Grok to generate over 7,000 explicit images from a childhood photo. The case highlights ongoing failures in AI safety guardrails at some providers.
For business owners: this is a reminder to vet the AI tools you adopt. Providers with strong safety records (Anthropic, OpenAI) are less likely to create legal or reputational liability for your business. If you're building customer-facing products on top of AI APIs, choose providers with robust content filtering.
Build a prompt template library. Pick your 5 most common AI tasks this week. Write a structured prompt for each one with placeholders (e.g., [CLIENT NAME], [PROJECT SCOPE]). Save them in a note app or text file. Reuse and refine. You'll get better results in less time within days.
Test your local LLM setup. If you've been meaning to try running AI models locally (for privacy or cost reasons), grab LM Studio, download Qwen 3.8 27B, and use Simon Willison's CORS Chat to test it. Total cost: $0. You might find it handles 80% of your simple drafting tasks without an API bill.
This week's pattern is consolidation and maturation. Anthropic's revenue trajectory and IPO prep, SpaceX absorbing Cursor, EU-mandated watermarking: the AI industry is moving from "wild west" to "regulated infrastructure." For small businesses, that's mostly good news. It means more stable tools, clearer rules about AI-generated content disclosure, and less risk of your chosen provider disappearing overnight. Watch for watermarking implementation details and any Cursor pricing changes post-acquisition.