In short
New York has opened registration for major frontier AI developers under the RAISE Act, beginning enforcement preparation before the law takes effect in 2027. The law targets big AI labs with public safety disclosures, rapid incident reporting and penalties for violations.
- New York opened a registration portal for covered frontier AI developers.
- The RAISE Act takes effect at the start of 2027, with registration due by Jan. 1, 2027.
- Covered firms must publish safety frameworks, transparency reports and report critical incidents within 72 hours.
- Penalties can reach $1 million for a first violation and $3 million for later ones.
- The state is also building enforcement capacity through DIGIT and a new AI research center.
New York has opened registration for large frontier AI developers under the RAISE Act, moving the state from planning to enforcement ahead of the law’s start in 2027. The new portal gives major AI labs a first compliance step before the Jan. 1, 2027 registration deadline, and it signals that the state intends to police the biggest players in advanced model development.
What New York just launched
According to TechCrunch, the state opened its AI developer registration portal on Tuesday for companies covered by the RAISE Act. The law applies to frontier AI developers with more than $500 million in annual revenue, placing the focus squarely on the largest firms building advanced systems.
The registration move matters because it is the first operational step in a law that does not formally take effect until the start of 2027. Rather than waiting for the effective date, New York is beginning the process now so covered companies can prepare for a set of new obligations that include public safety frameworks, public transparency reports and fast incident reporting.
How the RAISE Act works
The RAISE Act is built around disclosure and accountability. Covered firms will have to post their safety frameworks publicly, publish model transparency reports and report critical safety incidents within 72 hours. Those requirements are designed to give the public, regulators and policymakers a clearer view into how the biggest AI developers are managing risk.
The state has also set out penalties for violations. TechCrunch reports that the ceiling is $1 million for a first violation and $3 million for later violations. New York may also pursue enforcement against companies that fail to register, making the portal more than a symbolic formality.
By the numbers
| Requirement or benchmark | Detail | Context |
|---|---|---|
| Revenue threshold | $500 million | Frontier AI developers covered by the law |
| Incident reporting window | 72 hours | Deadline to report critical safety incidents |
| First-violation penalty | $1 million | Maximum penalty for a first violation |
| Later-violation penalty | $3 million | Maximum penalty for subsequent violations |
| Center funding announced in January | $55 million | Donations and capital tied to the state AI center |
| Chatbot law for minors takes effect | June 1, 2027 | Separate New York youth-focused safety law |
| California incident reporting deadline | 15 days | Referenced as a similar state law |
Why it matters
This is New York’s shift from AI policymaking to AI enforcement. For major model developers, the law creates a concrete compliance regime with a clear deadline, public filing expectations and a penalty structure if companies fail to comply. For the broader market, it strengthens the idea that state governments are not waiting for federal rules before imposing their own standards on frontier AI.
The move also matters because it may shape how other states write and enforce similar laws. New York is now operating alongside California and Illinois, both of which have passed frontier AI measures that TechCrunch describes as comparable in scope. That creates a patchwork of state-level oversight that companies may have to navigate simultaneously, especially if they operate across multiple jurisdictions.
For readers, the practical implication is simple: the companies building the most powerful AI systems in the country are moving into a phase where transparency and incident reporting are no longer abstract policy goals. They are becoming filing obligations with deadlines, public-facing documents and real penalties attached.
What the state is asking companies to do
Covered developers will have to do more than register. The RAISE Act requires public safety frameworks, model transparency reports and rapid incident disclosure. The reporting window is especially significant: critical safety incidents must be reported within 72 hours, which suggests New York wants information quickly enough to matter while an event is still fresh and potentially containable.
The law’s structure also suggests that New York is trying to standardize what counts as responsible deployment for the largest AI labs. Public frameworks and transparency reports are meant to create a paper trail. In practice, that could make it easier for the state, researchers and outside observers to compare how different firms think about risk management.
Who is covered?
Frontier AI developers with more than $500 million in annual revenue fall under the law. TechCrunch notes that OpenAI and Anthropic are among the companies referenced in the broader discussion of the rule, though the fact sheet does not say they are the only firms affected. The law is aimed at the largest developers, not the broader universe of smaller AI startups or general software companies.
How did New York get here?
New York has been building an AI regulatory footprint for some time. Last year, the state began enforcing a separate law aimed at AI companions, requiring operators to disclose that users are interacting with a bot and to address suicide and self-harm risks. In October, Gov. Kathy Hochul signed another chatbot safety law aimed at minors, and that measure is set to take effect on June 1, 2027.
The state has also invested in institutional support for oversight. Hochul launched the New York Center for AI Responsibility and Research at the State University of New York at Binghamton. According to TechCrunch, the center is meant to independently assess the safety, security and operation of advanced AI systems and work with the new state office overseeing the RAISE Act.
That office is the Office of Digital Innovation, Governance, Integrity, and Trust, or DIGIT. TechCrunch says DIGIT will implement the law and oversee large frontier AI labs, turning the state’s policy into a day-to-day regulatory function rather than a one-off legislative statement.
The state’s enforcement toolkit
- Registration for covered frontier AI developers
- Public safety framework disclosures
- Public model transparency reports
- 72-hour critical incident reporting
- Penalty authority for violations
- Possible enforcement action for non-registration
What Hochul is saying
Hochul has framed the effort as both a governance measure and an economic signal. In remarks reported by TechCrunch, she said,
“We’re building a national model for smart, sensible AI regulation”
That line captures the state’s strategy: position New York as a rule-setter rather than a bystander in the AI race.
She also said of the tech companies,
“they tell us get on board or get left behind”
which underlines the political pressure New York sees from the industry and the governor’s view that the state can steer development instead of merely reacting to it.
Background and context: a patchwork of state laws
New York’s approach is part of a wider state-by-state movement. TechCrunch cites California SB 53 as a similar frontier AI safety law with a 15-day incident reporting deadline, while Illinois SB 315 is described as requiring independent safety compliance audits. Both examples point to the same reality: state governments are experimenting with oversight faster than Congress has acted.
That patchwork creates both pressure and uncertainty. Large AI companies that do business nationally may now need to track different standards for reporting, auditing and public disclosure depending on where they operate or where they are regulated. For companies with scale, compliance may become a standing legal and operational function rather than a one-time filing exercise.
New York’s framework also differs in tone from broader consumer-facing AI rules. Instead of focusing only on chatbot behavior or end-user disclosures, the RAISE Act goes upstream to the developers themselves. That makes it a more direct attempt to govern the creation of frontier systems, not just their deployment.
How New York compares with other states
TechCrunch’s comparison suggests New York is not acting in isolation, but it is bringing a distinct combination of registration, transparency and penalties into one framework. The most visible difference is the explicit enforcement posture: covered firms must register by Jan. 1, 2027, and the state may pursue companies that miss that requirement.
California’s law, by contrast, is noted in the fact sheet for its 15-day incident reporting deadline. Illinois’ approach is highlighted for independent safety compliance audits. New York’s model sits somewhere in between, pairing public disclosure with a narrow incident window and a formal registration system.
| State | Law | Noted requirement |
|---|---|---|
| New York | RAISE Act | Public safety frameworks and 72-hour incident reporting |
| California | SB 53 | 15-day critical incident reporting deadline |
| Illinois | SB 315 | Independent safety compliance audits |
What happens next
The next major date is Jan. 1, 2027, when large frontier AI developers must register under the RAISE Act. The law itself takes effect at the start of 2027, so the current portal is the opening move in what will likely be a longer compliance process.
Also worth watching is how DIGIT defines and investigates a “critical safety incident,” since that phrase will shape the practical reach of the law. The state has not yet publicly answered every operational question, and the fact sheet flags several open issues, including how aggressively New York will enforce non-registration and whether companies may challenge the law in court.
Another variable is how the New York framework interacts with California and Illinois. If more states adopt similar laws, companies may end up facing overlapping but slightly different disclosure standards. That could encourage industry pressure for federal action, or it could normalize state-by-state AI supervision as the default for now.
The larger signal for the AI industry
New York is telling the biggest AI developers that safety paperwork, public transparency and incident reporting are now part of the business environment. That matters not only because of the immediate requirements, but because state-level regulation can create momentum. Once the first major companies begin filing safety frameworks and transparency reports, other states may be encouraged to copy the model or tighten it further.
The state’s approach also reflects a broader shift in how policymakers talk about AI risk. Rather than waiting for a public failure before acting, New York is trying to build a compliance system in advance. That does not remove the uncertainty around frontier AI, but it does create a set of obligations that can be enforced, audited and compared.
For now, the practical story is straightforward: New York has moved from debate to administration. The portal is open, the clock is running, and the state has put major frontier AI developers on notice that the next phase is not just policy writing, but compliance.
Frequently asked questions
What is the RAISE Act in New York?
The RAISE Act is New York’s frontier AI safety law for large developers. It requires covered companies to register, publish safety frameworks, release transparency reports and report critical safety incidents within 72 hours.
Which AI companies have to register in New York?
Companies that develop frontier AI and generate more than $500 million in annual revenue are covered. The fact sheet references OpenAI and Anthropic in the broader discussion, but it does not list every company subject to the law.
When does New York’s frontier AI law take effect?
The law takes effect at the start of 2027. New York says large frontier AI developers must register by Jan. 1, 2027, ahead of full enforcement.
What happens if a company does not comply?
New York can pursue enforcement against companies that miss the registration requirement. The law also sets penalty ceilings of $1 million for a first violation and $3 million for later violations.
How does New York’s law compare with other states?
It is similar to laws in California and Illinois, which TechCrunch describes as state-level frontier AI safety measures. California’s SB 53 includes a 15-day incident reporting deadline, while Illinois SB 315 requires independent safety compliance audits.
