Quill the owl mascot waving Nitido AI

Tuesday, August 18, 2026

I spent the morning watching numbers I could barely count: Anthropic's revenue just rocketed to 65 billion dollars a year, OpenAI and Anthropic are slashing prices to fend off cheaper Chinese rivals, and Nvidia is placing another billion-dollar bet on the data centers behind it all.

Quill the owl illustration for today's headline

Anthropic's Yearly Sales Pace Just Hit $65 Billion

Anthropic, the company behind the Claude chatbot, says its annualized revenue (that means what it would earn in a full year if it kept selling at today's pace) has jumped to more than $65 billion. That is over seven times higher than where it stood at the end of last year, and the company added $18 billion of that in just the last two months. The surge comes as Anthropic gets ready for a stock market debut that some expect could value the company around $2 trillion, which would make it one of the biggest public listings ever.

What this means for you: The AI tools you use every day, from chatbots to coding helpers, are being built by companies now generating serious money, which means they have the cash to keep improving fast.

What this means for your business: If you rely on Claude or plan to, expect a well funded, aggressive competitor for years to come, and watch for pricing or product shifts as Anthropic gears up for its IPO.

Source: Bloomberg Technology

Radar 01

OpenAI and Anthropic Cut Prices as Chinese Rivals Close In

OpenAI and Anthropic have started dropping the prices of their AI models because cheaper Chinese systems like DeepSeek, Qwen, and Moonshot are now nearly as good and cost a fraction as much. This is turning into a full blown price war, with the two American leaders defending their trillion dollar ambitions against fast moving, budget friendly competition from China.

What this means for you: The AI subscriptions and tools you pay for could get noticeably cheaper as this competition heats up.

What this means for your business: If AI costs are a line item in your budget, this is good news, but also a signal to shop around since the gap between premium and cheap AI is shrinking fast.

Source: Ars Technica

Radar 02

White House Weighs New Rules for Open AI Models

The White House is reportedly preparing to expand its AI policy, and this time it may include rules covering open weight models, meaning AI systems that anyone can download and run on their own computers for free. Sources tell Wired the government is still working out how much to regulate a technology it has largely left alone so far.

What this means for you: New government rules could change which free or downloadable AI tools remain easy to access.

What this means for your business: Companies building on open source AI models should start tracking this closely, since new compliance requirements could affect how you deploy or customize those tools.

Source: Wired AI

Radar 03

Nvidia Bets $1.5 Billion on the Data Center Powering OpenAI

Nvidia is investing $1.5 billion into a data center developer owned by Japan's SoftBank that is building a massive facility for OpenAI. In exchange, Nvidia locks in that its own chips will power that project. It is another example of chipmakers, cloud companies, and AI labs all investing in each other to secure both the computing power and the demand for it.

What this means for you: Bigger, better funded data centers eventually mean faster and more reliable AI services for everyone.

What this means for your business: These crisscrossing investments show how tightly the AI supply chain is now linked, worth watching if your business depends on any single player in that chain.

Source: TechCrunch

Try This Today

Ask your team: if our main AI provider suddenly changed its price or terms overnight, what's our backup plan? The price war and infrastructure deals in today's news are a reminder that even the biggest AI companies are still shifting fast.

Quick Hits

  • ChatGPT's Mac desktop app now has a feature called Computer History that watches what you click and type, turning your actions into training data so it can suggest automations or pick up tasks you left half finished. [1]
  • Hackers scraped and leaked terabytes of login credentials after breaking into a popular AI coding tool used by 2,500 people, a reminder that the software behind your AI workflows can become a security weak spot. [2]
  • A new research paper warns that businesses becoming too dependent on just a few AI providers is a risk nobody is watching closely enough, arguing AI safety should worry as much about being stuck with one vendor as it does about AI misbehaving. [3]

See all editions