Breaking — OpenAI CEO Sam Altman told staff in a Slack message that the company expects to go public “within the next year.” In the same week, OpenAI’s chief scientist revealed a new model codenamed 5.6 — a “meaningful improvement” over GPT-5.5 — launching this month. But Altman added a caveat that changes everything.
The Three Reveals
The RSI Caveat Is the Real News
Most coverage will lead with the IPO timeline or the new model. The structural story is the caveat.
RSI — Recursive Self-Improvement — is the scenario where AI becomes capable of improving its own capabilities autonomously. AI that creates better AI. Altman is telling his staff, plainly: if this happens, going public might not make sense, because the technology could change so fast that quarterly earnings reports become meaningless.
“The faster the potential RSI takeoff looks like it could be, the more it could be advantageous to delay an IPO. The technology and the world may change in surprising ways, and there might be good reasons to be a private company during that time.”
— Sam Altman, Slack message to OpenAI staff
Read that carefully. The CEO of an $852B company just told his employees that the company’s own technology might make public markets irrelevant.
This connects directly to what Dario Amodei wrote in his essay today: “Powerful AI” — a country of geniuses in a datacenter — within 1-2 years. Both CEOs of the two leading AI labs are now publicly discussing the same timeline for transformative AI capability.
Model 5.6: The Capability Race Continues
Altman’s Slack message landed the same day Anthropic released Fable 5 — which scored 80.3% on SWE-Bench Pro vs GPT-5.5’s 58.6%. A 22-percentage-point gap.
Model 5.6 is OpenAI’s response. If it is a “meaningful improvement” over 5.5, it needs to close that gap — or risk losing the frontier positioning that justifies an $852B valuation.
The timing: Fable 5 launches June 9. OpenAI announces 5.6 June 10. Launch “this month.” The capability race between Layer 4 providers has never been faster.
The Product Overhang Read
OpenAI is managing three overhangs simultaneously:
- Capital overhang: $122B raised privately, $852B valuation, investors need liquidity → S-1 filed, IPO within a year
- Capability overhang: Model 5.6 in the pipeline, “meaningful improvement” → launching this month to counter Fable 5
- Existential overhang: RSI — if it happens, the entire market structure changes → Altman explicitly tells staff this could delay the IPO
The first two overhangs have scheduled release dates. The third one doesn’t. And Altman is telling his team it might override everything else.
The IPO Race Updated
But Altman just introduced a variable none of the others have: the IPO might not happen if the technology advances too fast. No other CEO filing to go public has ever said that.
Related:
OpenAI S-1: What $852B Buys at Layer 4
OpenAI Three-Phase Plan
Dario Amodei: Policy on the AI Exponential
Claude Fable 5 Launch
Product Overhang Doctrine
Source: The Information (June 10, 2026), Sam Altman Slack message, Jakub Pachocki staff message
