New York opens registration portal for RAISE Act; large frontier AI developers must register by Jan 1, 2027

On 10 Nov 2026 New York opened a registration portal for large frontier AI developers, defined as those with over $500M annual revenue (OpenAI and Anthropic are the named examples). They must register by 1 Jan 2027, when the RAISE Act takes effect. The Act requires published safety frameworks, critical incident reports within 72 hours and public transparency reports. A new office, DIGIT, will oversee compliance, with penalties of up to $1M for a first violation and $3M for later ones. Gov. Hochul also launched a research centre at SUNY Binghamton to evaluate advanced AI systems and review safety frameworks. The article links to no primary document, so these details rest on the coverage and the Governor's statements.

This turns an existing law into operational duties with a deadline. The biggest AI labs now have about seven weeks to register and must build processes for 72-hour incident reporting and for publishing safety frameworks. The penalties are small against these companies' revenues, so the real pressure comes from public disclosure and reputational risk, not the fines. It also adds to a growing state-level patchwork (California SB 53 and Illinois SB 315 are mentioned), which could push firms to standardise on the strictest requirements or press for federal pre-emption. How much it changes behaviour depends on how DIGIT defines a "critical incident" and whether it actually enforces.

Hochul's "national model" and "communities above corporations" language is political framing, not evidence of impact, and "one of the country's most significant AI safety laws" is untested until there is enforcement. The RAISE Act was already law, so this is an implementation step, not a new rule. The coverage also gives no sign that any company has registered or objected, and it says little about how "critical incident" will be defined.

1. 01Which companies register before 1 Jan 2027, and whether any challenge the law or the $500M threshold in court or lobby for federal pre-emption.
2. 02DIGIT's published guidance or rules defining 'critical incident' and the format of safety frameworks and transparency reports, plus the first reports made public after 1 Jan 2027.
3. 03The first enforcement action or penalty, and whether the SUNY Binghamton centre publishes any actual evaluations or framework reviews in 2027.