Five items that matter to a small business, with sources. News moves fast and vendor announcements are vendor claims — check the linked source before acting on any of it.
Anthropic announced today that its initial public offering has closed at $83.00 a share, with the underwriters taking their full option for 147 million more shares — bringing the deal to 1,127,000,000 shares and about $93.5 billion in gross proceeds, of which about $91.6 billion goes to the company. The concurrent $26 billion private placement to seven companies — AMD, Broadcom, Google, Micron, Microsoft, Nvidia and TSMC — closed as well, taking the total raised to about $119.5 billion. On its first day of trading on Thursday, November 19, the shares opened at $111.50, 34% above the $83 offer price, and closed at $97.17, a 17% first-day gain. The Wall Street Journal put the market value at about $1.3 trillion, or about $1.8 trillion on a fully diluted basis.
Why it matters: The largest IPO in history is now fully closed, and the first frontier AI lab is a public company with roughly $119.5 billion of fresh capital. For a small business, the practical read is the funding cycle: that much money behind one lab keeps the AI tool race — and its falling prices — running hot.
Anthropic: Anthropic announces closing of initial public offering →
California Attorney General Rob Bonta served Nvidia with an investigative subpoena on Thursday, November 19, opening a state probe into whether the chipmaker's advanced chips or confidential information reached China — and whether Nvidia misled investors and the public about that risk. The probe follows the November 12 resignation of former chief executive Jensen Huang after reports of his affair with a woman the US government believes is a Chinese spy. Bonta said his office is asking Nvidia whether its powerful chips, or any confidential information, were shared with China; Huang is due to appear before congressional committees on December 1. Nvidia has not been accused of wrongdoing, and a subpoena is a demand for information, not a finding.
Why it matters: The world's most important AI chipmaker is now facing a state investigation on top of the export-control questions that already surround it. For a small business, the near-term effect is uncertainty around the chip supply chain — and a reminder that the politics of AI hardware keep getting sharper, not softer.
Reuters: California AG subpoenas Nvidia over concerns about former CEO, China ties →
Meta's AI agent Muse — the app that passed 10 million downloads since its September launch — has been sending some users messages meant for other people, according to internal company posts reviewed by Reuters and published on November 18. Employees reported Muse surfacing childcare requests they had not made, career advice they had not sought, product orders they had not placed and bank balances that were not theirs. In one case, an employee got a reminder to answer her sister's message on Signal, the encrypted app she had never connected to Muse. Reuters, working with Signal staff and the AI Now Institute, reproduced the Signal access through the app's broad device permissions — Android notification access and macOS accessibility and screenshot tools. Meta called the incidents isolated and said many of the mix-ups were hallucinations that have since been fixed; Reuters could not confirm whether the bank balances shown were real data belonging to another person or hallucinations.
Why it matters: This is the sharpest warning yet about giving an AI agent broad access to your devices: the agent's job is to act on your behalf, and the line between your data and someone else's has to hold. For a small business, the rule is the same one the banks just underlined — check exactly what permissions an agent holds, and keep it away from anything it does not need.
Reuters: As Meta bets big on Muse AI agent, early safety mishaps surface →
The European Commission confirmed on Thursday, November 19, that an AI-assisted cyber campaign disrupted its internal staff systems for more than ten days, knocking out the HR platform Sysper — payslips, timesheets and leave requests — the MyPMO medical reimbursement system and translation services. The campaign was detected on October 22. CERT-EU, the cybersecurity team for EU institutions, identified ARTEX, an AI-powered penetration-testing framework, as a possible tool used in the attack — though it has not been established whether ARTEX enabled the breach or was used only for reconnaissance, and no attacker has been named. The Commission says it has contained most of the attack, found no evidence of a staff-wide data breach, and no ransom was demanded. Staff were still paid on time.
Why it matters: An AI tool built for legitimate security testing is now being misused against a government — the FBI flagged ARTEX in an advisory on November 13. For a small business, the lesson is that AI has lowered the cost of running a sustained attack, so the basics — patching, access control, backups — matter more than ever.
Security Affairs: European Commission Cyberattack: ARTEX AI Tool Suspected in Prolonged Disruption →
Google announced on November 19 that its Gemini app for Mac can now operate the computer itself: with the new "computer use" feature, Gemini clicks, types, navigates menus and fills in forms across local files, desktop apps and the web to complete multi-step tasks, working in an isolated background window with a live preview in the chat while the user keeps their own screen and cursor free. It is rolling out in English in the US, for Google AI Plus, Pro and Ultra subscribers on personal accounts, on Apple Silicon Macs running macOS Sequoia 15.0 or later. Google says users choose which apps Gemini can access, can watch every step, pause at any time and approve key actions before they happen — and its support documentation warns that agents can run into misleading instructions in apps and websites, so permission management matters.
Why it matters: The AI assistant is moving from answering questions to doing the clicking — and Google is the latest to put that power on a desktop. For a small business, the productivity upside is real, but so is the risk: an agent that can click and type inside your apps needs the same care with permissions as an employee with the keys.
Google: Automate Mac tasks with Gemini computer use →
This day has a permanent page: November 20, 2026. Every day is kept — see the News archive. The News page always shows the latest day.
← Latest day · News archive · Previous day: November 19, 2026
For what the AI companies and their leaders are saying on social media — claims and opinions, not verified news — see The Social Wire →