AI News Today September 16 - OpenAI Wants $1.5 Trillion, AI Giants Caught Talking In Secret, China Hands Out A Free Super-Agent

Four days ago the people running the largest AI companies asked the industry to slow down. Today one of them is trying to raise money at almost double the price it fetched in March.

OpenAI is in talks over a funding round that would value it somewhere between $1.2 trillion and $1.5 trillion, a number that matters well beyond Silicon Valley because these private valuations are the foundation the entire public AI trade is priced against.

At the same time, three rival labs have confirmed they have been quietly coordinating for weeks, and a Chinese laboratory has given away a frontier research agent for nothing.

Here's everything you need to know today.

Why Does OpenAI Want $1.5 Trillion?

Six months ago, OpenAI was the most valuable private company in history. It is now asking the market to nearly double that.

Investors approached the company offering to invest at $1.2 trillion, and OpenAI is reportedly holding out for $1.5 trillion instead.

The reason is simple arithmetic. OpenAI spent $34 billion last year training models, annualised revenue passed $40 billion last month, and Sam Altman has just ruled out the stock market listing that would have funded the next round of spending. A company burning money at that rate has to raise it somewhere.

This is why it matters to anyone who has never bought a technology stock in their life. OpenAI is private, so there is no share price to argue with, which means every public AI company gets valued partly against a number set behind closed doors by a few dozen funds. If that number is wrong, a great deal of retirement money is wrong with it.

What Are OpenAI, Anthropic and Google Doing Together?

The three companies that dominate AI in America have just admitted they have been talking to each other privately, and they did not ask permission first.

OpenAI's policy chief confirmed the company has spent weeks in safety talks with Anthropic and Google DeepMind, and said no antitrust waiver was needed to hold them.

That last part is the problem. Three firms controlling most of a market, agreeing between themselves how fast to move, is the textbook shape of an antitrust case, which is exactly why Anthropic asked Washington for legal cover before it started.

Critics have a blunter reading, which is that slowing down happens to be cheapest for whoever is already in front.

Either way, the companies deciding how fast AI develops are now doing it in a room with no regulator in it.

China Just Released a Free Frontier Agent

While Washington and Silicon Valley argued about brakes, Beijing quietly did the opposite.

Shanghai AI Laboratory published a 744 billion parameter research agent under an open licence, with no announcement and no price attached. Anyone can download it and run it themselves, and its own benchmark scores claim it edges out OpenAI's and Anthropic's flagship models on web research tasks, though no independent lab has verified that yet.

The commercial point is sharper than the technical one. American labs sell access to their best models for thousands of dollars a month, and a Chinese state laboratory has just given a comparable tool away for free.

It is also the single strongest argument against an American slowdown, and the one Trump and House Speaker Mike Johnson have both reached for this week.

The Last People to Find Out Are the Ones Who Pay for It

OpenAI went from $300 billion to $1.5 trillion in eighteen months. You were never offered a single share.

That is the part worth worrying about.

Enormous sums are moving into a handful of technologies right now, and the public always hears about it late, once the easy money has already been made by someone else.

The $5 Trillion Signal is our free report on the five technologies taking in that money. What they are, who is funding them, and what to watch for, in plain English.

Get your free copy of The $5 Trillion Signal below.