Home / Company
Company

Everything an AI platform team would build. Delivered as a product.

Every company's AI stack is three layers: models, rules, and the applications employees open. Enterprises with a platform team weld those layers into something that fits them. We exist for everyone else.

free·hold/ˈfriː.hoʊld/ · noun · property law An estate held outright, free of any landlord. The opposite of a leasehold: occupancy on the owner's terms, reverting to the owner when the lease ends.

Why the name.

Renting AI seat by seat means living on the landlord's terms: their prices, their data-handling clauses, their model retirements, their schedule. A freehold is the other tenure: what you hold, you hold outright. Your hardware on your balance sheet, your tenancy in your name, your keys, your data inside your perimeter. Even when a workload runs against a rented API, it runs under your governance. And what accumulates over time (the corpus, the memory, the working habits of your teams) accumulates to you.

Principles we build against.

Each one is checkable in the product.

Honesty in numbers

Numbers carry their provenance

Every number in the console is labelled measured, derived, or declared, and what we can't measure shows as a labelled absence. The same rule governs this website.

True names

You're trusted to know

Models and applications appear under their real names, and adopted open source is labelled as adopted open source, with licenses honored and notices preserved.

A visible catalog

Priced on the shelf

Transparent per-app pricing, a deployed product that is complete as delivered, and new capability appearing in the open, on the shelf.

Architecture over promises

Sovereignty you can test

"Your data stays yours" is enforced by system structure and validated by injected-fault tests before we claim it anywhere.

One accountable party

Your recourse is always us

Hardware vendors, platform vendors, and upstream projects all stand behind us, as our subcontractors and suppliers. You hold one SLA, and it has our name on it.

Compose, don't rebuild

Engineering where it counts

Model serving and the certified stack are solved, so we build on NVIDIA AI Enterprise as it ships. Our engineering goes to the layer left open: the finished catalog and the governance experience above it.

The company

Built by people who've shipped this before.

Freehold Technologies, Inc. is a Delaware C-corporation founded in 2026. The founding team met building and delivering enterprise AI inside a global consultancy's GenAI practice, shipping AI products to Fortune 500 clients and running the delivery model behind them, then went deeper into technical roles at enterprise AI companies before starting Freehold.

The pattern we kept watching: the companies that got real value from AI all had a platform team welding the stack together, and everyone else got seat licenses. Freehold is the product version of that missing team.

Where we are

Early, and honest about it.

The governance core is built and fault-tested: the config-only console, the governed gateway with per-person budgets and the content-free audit, identity federation, and third-party providers governed under customer keys. The full console experience is designed against its stated principles, and the deployment pipeline is the current engineering front.

We're a design-partner company right now, on purpose: a small number of deployments, worked closely, steering the catalog. That's the honest place to meet us.

The design partner program.

We're selecting a small number of early deployment partners, with priority for regulated industries and teams that will push the catalog hardest.

What partners get
  • Design-partner terms, agreed in full before you sign
  • Direct access to the people building the product
  • Catalog priorities steered by your workloads and your vertical
  • The same sovereignty architecture and SLA discipline as any future customer, from day one
What we ask
  • Real workloads and real users, from the first month
  • A named IT owner who'll run the console and tell us where it fails them
  • Candor: the product gets better at the rate partners are blunt
  • A reference conversation when, and only when, the deployment has earned it
For app founders

Your app, enterprise-ready, in our catalog.

If you've built an AI application buyers want and your enterprise deals stall on governance questions (where does our data go, can this run in our VPC or on our hardware, how do we control spend, who audits usage, how does it plug into our identity provider), catalog placement answers all of it at once.

We integrate your product under the app contract, and it ships with governance, cost control, and every deployment posture already solved, with you at the table as the vendor behind the code from day one.

Talk about catalog placement →
For the channel

Resellers and MSPs who own the hardware relationship.

If you sell or manage the certified hardware lines and your customers are asking for the AI layer you can't deliver alone, our layer rides on exactly the stack you already ship. The operating relationship stays coherent: one SLA, clean subcontracting lines, and no ambiguity about who the customer calls.

Start a channel conversation →
Contact

One inbox, read by the founders.

Deployments, partnerships, security questions, or a hole you think you've found in our story: all of it lands in the same place, and all of it gets a straight answer.