Illustration for: Two Weeks After 'Pace The Frontier,' Nothing Slowed

Two Weeks After 'Pace The Frontier,' Nothing Slowed

Two weeks after Dario Amodei's essay calling on AI labs to slow down together, the same three labs are still shipping at full speed -- and now face a federal antitrust suit alleging the essay itself was evidence of collusion.

By the Numbers

Sep 12, 2026
Essay published
Reported Sep 23
Standards-body talks
Sep 18, 2026
Antitrust suit filed
4 paid subscribers
Plaintiffs
Sherman Act, Sec. 1
Claim
TC
By the Markets Desk
Edited by Trace Cohen · Early-stage VC & angel · Founder, New York Venture Partners
2 min read
ShareXLinkedInEmail

THE RUNDOWN

1

The September 12 'pace the frontier' essay produced two concrete follow-ons: a Frontier AI Standards Agency effort courting Sriram Krishnan to run it, and, on September 18, a federal antitrust lawsuit alleging the public coordination itself violates the Sherman Act.

2

None of Amodei, Altman or Hassabis has publicly committed to slowing any specific release -- GPT-6 pricing cuts and Anthropic's newest model tier both shipped in the two weeks since the essay published.

3

The standards-agency talks and the antitrust suit are two sides of the same fact pattern: regulators-in-waiting and plaintiffs' lawyers are both treating public coordination among competitors as evidence, whatever the labs intended by it.

4

For founders and LPs with exposure to any of these labs, the real signal isn't the essay's rhetoric -- it's whether release cadence actually changes, and two weeks of data can't yet show that either way.

TC

The VC Read · Trace's Take

Trace Cohen

The lawsuit is the more interesting development than the essay itself: four paying subscribers just turned a PR gesture into a discovery fight, and discovery is where 'we agreed to pace together' either becomes emails proving coordination or nothing at all. Diligence item for anyone with cap table exposure to these labs: watch whether the standards-agency talks produce an actual documented governance structure -- bylaws, funding, enforcement -- in the next quarter, or quietly die the way most 'voluntary industry body' proposals do once formalizing it creates its own antitrust exposure.

Analysis

It has been two weeks since Anthropic CEO Dario Amodei published his September 12 essay urging AI labs to "pace the frontier" -- explicit industry cooperation on decelerating capability gains in favor of safety work. OpenAI's Sam Altman, xAI's Elon Musk and Google DeepMind's Demis Hassabis each publicly endorsed the idea within hours of publication, a level of same-day, on-the-record agreement among direct competitors that is unusual even by AI-industry standards.

Two real developments have followed. Google, OpenAI and Anthropic are now reportedly forming a Frontier AI Standards Agency and have approached Sriram Krishnan, the former White House AI policy adviser, to run it -- a voluntary, FINRA-style testing and safety body meant to operate ahead of any binding legislation. Separately, four paying subscribers to ChatGPT, Claude, Grok and Gemini filed a federal antitrust suit in the Northern District of California on September 18, arguing the public coordination around Amodei's essay amounts to an illegal restraint of trade under the Sherman Act, according to CBS News.

The two threads point in opposite directions. The standards-agency effort is the labs trying to get ahead of regulation by writing their own rules first, the same playbook FINRA itself grew out of in securities markets. The antitrust suit takes the opposite reading of the identical facts: that any public agreement among direct competitors to slow a dimension of competition, safety-motivated or not, is exactly the kind of coordination Section 1 of the Sherman Act exists to catch, regardless of the labs' stated intent.

“What hasn't happened in the two weeks between is any observable slowdown.”

What hasn't happened in the two weeks between is any observable slowdown. OpenAI's GPT-6 pricing moved sharply lower in the same window, and Anthropic shipped a new pricing tier of its own -- both moves that read as competitive intensity, not restraint. That's the risk for anyone taking the essay at face value: a voluntary pledge with no enforcement mechanism and an obvious first-mover disadvantage for whichever lab actually slows down first was always more likely to produce a standards-body press release than a changed release calendar.

The counterweight worth stating plainly: nothing here proves coordination beyond three CEOs publicly agreeing with an essay, and "agreeing in public" is a long way from an enforceable pact in antitrust law. The lawsuit's four plaintiffs will need discovery -- internal communications, not public statements -- to make a Sherman Act claim stick, and public agreement with a safety essay is the kind of thing labs might plausibly do independently, for genuinely different reasons, without ever coordinating.

The antitrust case's first substantive hearing, expected in the coming months, will test whether "we agreed to go slower together" becomes a legally cognizable restraint of trade or stays exactly what it looks like today: three CEOs agreeing publicly with an essay, then continuing to ship as fast as they can.

ShareXLinkedInEmail

Key Sources

2 sources

THE WIRE in your inbox— Tech, startup & VC news with Trace's take. Free, no spam.