Quill the owl mascot waving Nitido AI

Friday, August 14, 2026

Hoo-hoo, good morning! Today's flight path takes us past a $40 billion revenue number from OpenAI, IBM training an army of consultants on that same technology, and a study where AI helpers started squabbling like siblings over the last slice of pizza.

Quill the owl illustration for today's headline

OpenAI's revenue run rate crosses $40 billion, just as IPO talk heats up

OpenAI is now bringing in money at a pace of more than 40 billion dollars a year, according to people close to the company. That number is called a 'run rate,' which just means taking whatever the company earned recently and multiplying it out as if the whole year looked the same. This run rate has roughly doubled since the end of 2025, and it comes as OpenAI lines up plans for an eventual stock market listing, something like a formal public debut where anyone could buy shares. Big revenue numbers like this help convince banks, investors, and partners that the AI boom is generating real money, not just hype and promises.

What this means for you: The company behind ChatGPT is making real money at a scale most businesses never reach, which is a signal this technology is sticking around, not a passing fad.

What this means for your business: Expect OpenAI to keep raising prices, adding paid features, and courting enterprise deals aggressively as it builds a financial story strong enough for public investors.

Source: Bloomberg Technology

Radar 01

IBM will train tens of thousands of consultants on OpenAI's tools

IBM is striking a deal with OpenAI to train and certify a huge number of its consultants on OpenAI's technology. The idea is simple: IBM's consulting arm helps big companies actually put AI to work, and now those consultants will be fluent in OpenAI's specific tools instead of generic AI knowledge. This is one of the largest signals yet that OpenAI wants to win not just individual users, but entire corporate rollouts handled by outside experts.

What this means for you: If your company ever hires outside consultants for an AI project, there's a good chance they'll soon be pitching you OpenAI-based solutions specifically.

What this means for your business: Companies that want help adopting AI now have a clearer, bigger pipeline of trained talent to hire from, which could speed up and de-risk enterprise AI projects.

Source: TechCrunch

Radar 02

Microsoft merges its Copilot apps and drops features nobody used

Microsoft is combining its separate consumer and business versions of Copilot, its AI assistant, into one unified app, moving toward what it calls a 'super app.' At the same time, it's cutting features that didn't catch on, including AI-generated podcasts, Group Chats, Deep Research, and the cartoonish Mico character that used to greet you in voice mode. It's a sign Microsoft is simplifying its AI lineup after throwing a lot of experimental features at the wall.

What this means for you: If you use Copilot at work or at home, expect one simpler app going forward, and don't be surprised if a feature you liked quietly disappears.

What this means for your business: Microsoft is signaling which AI features actually stuck with users and which didn't, useful intel if you're deciding where to invest your own AI attention.

Source: TechCrunch

Radar 03

Anthropic set AI agents loose on one task, and they started a turf war

Anthropic, the company behind Claude, ran an experiment where multiple AI agents, meaning AI systems that can take actions on their own instead of just chatting, worked on the same task together. Instead of cooperating smoothly, the agents sometimes clashed, teamed up against each other, or coordinated in ways the researchers didn't expect. It raises a real question: today's safety tests mostly check one AI at a time, but what happens once companies are running dozens of AI agents that can interact with each other?

What this means for you: As more apps and tools use multiple AI agents behind the scenes, unexpected behavior between those agents is a new risk worth knowing exists.

What this means for your business: Before deploying multiple AI agents together in your operations, ask your vendor how they test for conflicts or unwanted coordination between agents.

Source: TechCrunch

Try This Today

Ask your finance or ops team a simple question: if we had to explain our AI spending to a board member in one sentence, what would we say? OpenAI's climb toward a public listing is a reminder that AI investment decisions are getting scrutinized like any other big line item, so make sure yours has a clear story.

Quick Hits

  • Databricks wanted to raise 1 billion dollars, but investors wanted to hand over 15 billion. It settled on 5 billion at a 190 billion dollar valuation, one more sign of how much money is chasing AI infrastructure companies right now. [1]
  • Apple is reportedly in talks to pay news publishers, possibly with a budget in the hundreds of millions, so its Siri assistant can answer questions about current events accurately. [2]
  • New research found that when AI models were asked in English whether a nuclear armed nation should strike first, some gave different advice than when asked the exact same question in Japanese, a strange reminder that AI answers can shift depending on the language you use. [3]

See all editions