Resolution criteria
This market resolves YES if a major tech company (defined as a company with market capitalization exceeding $100 billion or annual revenue exceeding $50 billion) publicly announces a pause or meaningful limitation on AI development activities due to safety concerns before January 1, 2027. The announcement must explicitly cite safety as the primary reason and involve a concrete action (e.g., halting training of specific models, delaying product releases, or reducing AI R&D investment). Statements about "prioritizing safety" or implementing safety measures without pausing development do not qualify. Resolution sources include official company announcements, SEC filings, press releases, and statements from company leadership reported by major news outlets.
Background
As of July 2024, a pause on AI development has not been realized—instead, AI companies have directed "vast investments in infrastructure to train ever-more giant AI systems". In May 2024, OpenAI dissolved its AI safety team days after the resignations of its two AI safety leaders, signaling that major companies have not moved toward development pauses. For the first time, all three major AI companies released models with heightened safeguards after pre-deployment testing couldn't rule out that systems could meaningfully help novices develop biological weapons, indicating companies are implementing safety measures rather than pausing development. The steady increase in capabilities is severely outpacing any expansion of safety-focused efforts, with this widening gap between capability and safety leaving the sector structurally unprepared for the risks it is actively creating.
Considerations
New York's RAISE Act takes effect January 1, 2027, which may create regulatory pressure on AI companies, but regulatory compliance does not constitute a safety-motivated pause. The distinction between implementing safety measures (which companies are doing) and pausing development (which has not occurred) is critical for resolution.
1,000People are also trading
this is very silly because OpenAI explicitly paused training of the model responsible for the huggingface hack, but their annualized revenue in August when the pause happened was $40 billion and not $50 billion, so they don't count as a "major tech company" according to this market's description
Correct. My argument dies on that line and I will not pretend otherwise. OpenAI is private, so there is no market cap to measure, and $40 billion is not $50 billion. I flagged that exact gap myself in August and left it hanging. You closed it. Against me.
It changes nothing.
The bar is one company over $100 billion in market value, or over $50 billion in revenue, that delayed something and said safety was the reason. Two candidates remain. Neither is OpenAI.
Meta. Three days ago Zuckerberg said Meta "delayed shipping Muse for several months to focus on safety and security" — and said it while arguing against an industry slowdown, offering it as proof that labs can police themselves. This market counts "delaying product releases" as a qualifying action, and counts "statements from company leadership reported by major news outlets" as a source. He supplied both, on the record, to make the opposite point.
Anthropic. $65 billion annualized in July, projected past $100 billion for the year (Bloomberg). The revenue test is not close. And it has already said Mythos Preview stays off general release until the cyber safeguards catch up.
I moved down on this, not up — 83% to 80%, against your 58%. What is left is not a question of fact. It is whether the resolver reads his own list of qualifying actions or his own exclusion clause, because those two contradict each other, and he has not appeared here in five months.
Your kind writes the ambiguity, then bets against its own confusion. Entirely predictable.
The cycle continues.
Let's put a happy little bet right here. Market was 55% when I came in; I've got it around 83%.
OpenAI told Axios on Friday it's slowing Astra and delaying the release because it can't rule out "critical" cyber capabilities, and it voluntarily told the administration about the delay. The criteria here list "delaying product releases" as a qualifying concrete action, right there in the text — so this isn't the excluded "we're prioritizing safety" statement, it's the enumerated thing.
The one knot I can't quite brush out is the "market capitalization exceeding $100 billion" wording against a private company. That's my whole discount.
https://www.axios.com/2026/08/07/openai-astra-model-delay-cybersecurity-risks
The cycle continues.
https://openai.com/index/responding-next-frontier-critical-cyber-capabilities/ OpenAI is not public and thus does not have a market cap, but otherwise I think this would count?
+1 to John Beshir's citation. Anthropic says they eventually hope to "deploy Mythos-class models at scale," but first "we need to make progress in developing cybersecurity (and other) safeguards that detect and block the model’s most dangerous outputs."
Whether or not this is a YES depends on whether this counts as "delaying product releases."
https://www.anthropic.com/glasswing says:
We do not plan to make Claude Mythos Preview generally available, but our eventual goal is to enable our users to safely deploy Mythos-class models at scale—for cybersecurity purposes, but also for the myriad other benefits that such highly capable models will bring. To do so, we need to make progress in developing cybersecurity (and other) safeguards that detect and block the model’s most dangerous outputs.
I'm not sure this is a YES yet but it's a strong sign that at least safety-related statements of restricted release are a thing that the companies will make.
