GREYMARKRESEARCH

Projects / Greymark Intelligence

Intelligence on the AI risks that resist oversight.

Greymark Intelligence maps the parts of the AI landscape the open record misses: open-source models and their derivatives, anonymous inference, crypto payment rails, grey markets, and the tools used to strip safety constraints from public models. We turn that into subscription intelligence for the organisations that have to price, police, or screen the resulting risk.

What we map

Coverage across the parts of the AI economy that operate outside conventional oversight.

Open-source models

Releases, fine-tunes, and redistribution, and how capability moves once a model is in the open.

The abliteration ecosystem

The tools and markets that strip safety constraints from public models and put the result back into circulation.

Anonymous inference

Commercial platforms that run models for paying users without attribution or logging.

Crypto rails

The payment and settlement layer beneath grey AI services.

Grey markets

Where access, compute, and capability are traded outside formal oversight.

Who it's for

Built for the organisations on the oversight side of this risk: the ones who price it, police it, or screen for it.

Insurance

Underwriters pricing AI-related exposure who need the risk mapped.

Trust and safety

Platform teams tracking how harmful capability is built, shared, and deployed.

KYC and compliance

Screening and monitoring functions that need defensible intelligence on services and counterparties.

The product

Greymark Intelligence is a subscription. Access is institutional, priced per organisation, and delivered as regular intelligence briefings with deeper assessments on the developments that warrant them. A founding cohort is forming now.

Every finding is owned by a named analyst and built from open sources, so it stands up to scrutiny.

Why it holds up

The value of this intelligence depends on it being defensible, so the rules are fixed. We are independent of the AI labs whose platforms we map, which means our findings cannot be bought. The product carries no identifiable safeguarding evidence or child-protection concerns. The operators we map cannot buy the intelligence that maps them. And while we name public actors where the evidence warrants it, we never expose information a person has kept private.

Read our independence statement

Request access

Greymark Intelligence is forming its founding cohort. To discuss whether it fits what your team needs to see, get in touch.

Get in touch