Suprvisr TLDR - June 1 2026

Starting this week, Cumulus lets you run agents directly inside your chat. No separate dashboard, no jumping between tools. You describe the task in the conversation, the agent does it, and it reports back in the same thread. It's the difference between asking for directions and having someone drive.

No credit card required.

Anthropic spent the week doing something its founders probably pictured back when they walked out of OpenAI in 2021: passing it. The maker of Claude raised $65 billion in a new round that values the company at $965 billion, ahead of OpenAI, which was last valued at $852 billion in March. Anthropic's number has more than doubled since February, when it sat at $380 billion. The company says its run-rate revenue crossed $47 billion earlier this month.

That growth isn't coming from chatbots. It's coming from companies. Claude Code has become one of the more widely used tools for getting software written, and the enterprise deals stacking up behind it (the kind where a single firm puts thousands of staff on the platform) are what turned the revenue line nearly vertical. The valuation is following the customers, not the other way around.

Why it matters: For two years the working assumption was that OpenAI was the company and everyone else was chasing. That framing is now harder to defend. More practically, both companies are expected to go public this fall, which means the stock market, not a handful of private investors, is about to set the price.

That's Anthropic's reported share of what U.S. companies spend across it and OpenAI combined, up from about 10% in early 2025, according to spend data from Ramp. The valuation flip this week is following real customer dollars, which is harder to argue with than investor sentiment.

Agents that can use any app you can. Microsoft made computer-using agents generally available in Copilot Studio. An agent can now operate any app a person can, with nothing built to connect it first. It reads the screen and clicks through the interface the way a human would, instead of relying on brittle scripts that snap the moment the software moves a button. If your team runs on older tools with no clean way in, that's the part to watch. An agent loose in your live systems is only as trustworthy as its guardrails, and most teams figure that out after they've turned it loose.

Siri's borrowed brain. Bloomberg got an early look at Apple's revamped Siri, the centerpiece of iOS 27 and the focus of its June 8 developer conference. The new assistant is built on a custom version of Google's Gemini, under a deal reportedly worth around a billion dollars a year, and behaves far more like a full chatbot. After two years of looking a step behind on AI, Apple is finally making its move, even if it needed a rival to do it.

Beijing keeps its talent close. China has begun requiring some of its top AI researchers and executives at firms like Alibaba and DeepSeek to get government approval before traveling abroad. Bloomberg reports the rules target people working on strategically important AI, and extending this kind of control into the private sector is new. If the strongest open models keep coming out of Chinese labs, it's worth noticing how tightly Beijing now guards the people building them.

It's worth knowing where your team's AI actually lives. Most companies still treat it as a chat window someone opens, asks a question, and closes. The common thread in this week's news, from Microsoft's agents to our own, is agents that stay in the conversation and do the task instead of just describing it. Pick one repetitive process your team handles by hand, something with clear steps, and try handing it to an agent start to finish. If you want help choosing the process worth starting with, that's exactly what we do over at Suprvisr AI.

Two of the three biggest names in AI are about to have share prices. We'll read the filings so you don't have to. See you next week.