Labs, researchers, medical teams and engineers are training new models to make discoveries. Through ABURA, companies lease, license or sell their data to them, with many buyers competing at once. Lease or license it, and you keep your data and the ownership.
What AI buyers have asked us for, alongside the data the world's biggest challenges need.
What a company's data can be worth to AI buyers. No ceiling.
No cost to you. The buyer pays.Introduce a company that holds data, and earn 10% of every deal it makes. No cap.
You never touch the data.The next generation of models is being trained to help solve the world's challenges for people and planet, find new medicines, build better infrastructure and tell new stories. To get there, they need data that was never published: the work companies do every day.
ABURA is a conscious data company. Every dataset is licensed with its owner's consent, personal data stays out, owners see every offer, and we champion data that helps people and planet.
Data on climate, nature, health access and social challenges, to train models that help solve the problems that matter most.
De-identified clinical, imaging and trial data, to train models that help spot disease earlier and bring new treatments sooner.
Drawings and site records with every revision, to train models that help design and build the cities, bridges and energy systems of the future.
Code written by people, with its full history, to train models that write, test and repair software.
Performance, tracking and match data, to train models that coach athletes, prevent injuries and bring fans closer to the game.
Footage, scripts, sound and archives, licensed properly, to train models that help tell the next generation of stories.
Research and sensor data, to train models that run experiments, factories and fleets.
Anonymised transaction data, to train models that catch fraud and widen fair access to credit.
Lease or license your data, and the buyer gets a copy to train on. The original stays with you, and so does the ownership. You decide who gets it, on what terms and for how long, and the same data can earn again with the next buyer. If you would rather make a clean break, you can sell it outright instead.
A company comes to us directly, or a scout who knows it makes the introduction.
Who owns it, what can be licensed and what stays private. Personal data stays out.
AI labs, research institutions, healthcare companies, financial institutions and more see the opportunity at once and make offers. You see every one, and you choose.
Straight from you to them. ABURA never holds your data, and with a lease or a licence the original never leaves you.
Hover over or tap a circle to follow its path.
Nine kinds of company data buyers can use, from companies trading today, winding down, or closed years ago.
Years of field data on climate, health, farming, education and inclusion, from companies working towards the Sustainable Development Goals.
Written by people before 2023, with commits, pull requests and tickets. COBOL, Rust, Go, Swift, MATLAB, C and C++ first.
Construction, infrastructure, machinery and industrial drawings, with every revision from first sketch to as built.
Contracts with their redlines, procedures, tickets and anonymised internal conversations.
Anonymised transaction, merchant and payment-flow data.
De-identified, consented clinical, imaging and trial data. Patient-identifiable records stay out.
Telemetry from machines, vehicles, fleets and sites.
Private research datasets, lab results and image libraries.
The archive outlives the company. Its former owners or a liquidator can license it.
Behind every impact company are years of blood, sweat and tears: the research, the field data and the discoveries gathered while trying to solve real problems for people and planet. When the funding runs out, that work too often disappears.
Through ABURA it can live on. Licensed to the teams training tomorrow's models, it keeps the mission going, returns value to founders and investors, and lets humanity build on what you started.
Most buyers name their own price and buy once. On ABURA, many buyers compete for a licence, you see every offer, and you keep the data.
A buyer pays for access for a set period. When it ends, you can lease it again.
A buyer gets a copy to train on, under your terms. The original and the ownership stay with you.
One payment, and the data becomes the buyer's. A clean break, if that suits you best.
Most owners lease or license, so the same data can earn again. You can decide once you see the offers.
One focused dataset, such as a codebase or a project archive.
A large, multi-year operational or engineering archive.
Rare data at scale that no one else can produce.
The largest data licences run to tens of millions a year.
Figures are indicative, drawn from publicly reported AI data deals. They are not a quote or an offer.
A house can be valued, insured and borrowed against. Company data now can too, so you can raise money without selling your data or giving up shares.
Earn 10% of every deal you introduce. No cap.
Your share is 10% of the first year's value of each deal. It comes from ABURA's share, never the owner's, once the owner has been paid.
A one-minute registration opens the field guide and the lead check.
When someone mentions old code, drawings or an archive, run the guided check. It tells you whether the lead is worth a conversation.
With the owner's agreement, send the message the check writes for you. We handle everything after that.
How long the buyers we talk to expect this window to stay open.
Lease or license, and buyers get a copy. You keep the original and the ownership.
ABURA never holds your data. A copy goes straight from you to the buyer.
AI labs need data to train their models that the internet simply cannot provide. Much of it sits in ordinary companies, including many that no longer trade. Spot it, check it here in a few minutes, and pass it on.
AI labs have already learned from the public internet.
What they need now was never published: code written by people, drawings revised by engineers, the inner workings of real businesses.
It sits in companies you know. Some of them closed years ago.
How long the buyers we talk to expect this window to stay open.
Companies that stopped trading, even years ago, are often the strongest leads.
You never touch the data. It stays with its owner throughout.
When a company winds up, its code, drawings and records usually sit untouched on a server or in a backup. The founders have moved on, and nobody thinks of it as an asset. Buyers do. Every stage of a company's life can qualify, and each calls for a slightly different question.
Ask what history they keep, and who would sign off on a conversation.
Often the best moment. The data is intact and the people who understand it are still there.
Ask who holds the files now: the founders, a liquidator, or a backup somewhere.
Older can be better. Code written before 2023 is exactly what buyers want, if the archive survived.
For a founder who has moved on, or a liquidator closing the books, the data may be the last thing of value the company has. Most of them have never been asked about it.
Each comes from what buyers have asked us for directly, so the bar is theirs.
The industry it served does not matter. The language does, and so does the history kept alongside it.
Old code, written by people, is the hottest data in AI right now.
History like this is part of what buyers value: the code, and the record of people building it.
Buyers want the journey of a design as well as where it ended. Construction, infrastructure, manufacturing and industrial plant come first, along with the design imagery these firms produce in volume.
Buyers want every revision, the revision cloud included, as well as the final sheet.
The procedures, knowledge, projects, tickets, decisions and conversations of a working company. Everything is anonymised before anything leaves it.
Buyers want years of weeks like this, anonymised, so AI learns how work is really done.
Legal and accounting documents with every draft kept, from first version to signed.
Lab, field and specialist datasets and research archives that are not public.
Archives that have never been online, with clear rights over every file.
Customer records, transactions, pricing and product history, operations, designs. It is all company data, and most of it has never been valued. A sister company of ABURA values company data independently, so its owner can license it, borrow against it, or both.
To AI labs and the data platforms that supply them, on terms the owner agrees.
Once valued and insured, company data can be taken by a lender as collateral, the way a bank lends against a home.
A friend, client or old colleague mentions code, drawings or operational data sitting unused, often at a company that has closed.
A short guided check, one question at a time. It tells you whether the lead is worth a conversation. It never gives a price.
Once they are happy to be introduced, the check writes the message for you. Send it to whoever invited you.
We speak with the owner in confidence, confirm what they own and introduce it to the right buyers. A sample set is needed only at the point of close, and the owner decides at every step.
Share a company's name only once the owner is happy to be introduced. Until then, the check keeps the name off the message for you.
No files, logins or screenshots. The data stays with its owner from the first conversation to the last.
Value is settled later, with the owner and the buyer. A number given early becomes the number everyone remembers.
If the code or drawings were built for a client, the client may own them. For a closed company, ask who holds the rights now. It is the question that matters most.
Health records, biometrics, card details and children's data are never part of it. Everything else is anonymised before it moves.
Very often, yes. A closed company's code, drawings and records usually still exist with the founders, a liquidator or in a backup, and whoever holds the rights can still license them.
Yes. A sister company of ABURA values company data independently, so it can be licensed or used as collateral, much as a bank lends against a valued, insured house. Owners can start a preliminary valuation on the valuation page, which opens without registration.
No. Rough answers are fine, and almost every question in the check has a "not sure" option. We work out the detail with the owner.
Never. It stays with its owner. You pass on a short message, and that is all.
Nothing at the start. A sample set is needed only at the point of close, and the owner decides what goes into it.
AI labs and the data platforms that supply them. We introduce owners to them in confidence, and only with the owner's agreement.
Something like: "AI labs are looking for data that has never been published, and what your company built may be exactly that. Would you be open to a short, confidential conversation to see if it qualifies? Nothing is shared without your agreement." The check writes a version for each lead, matched to what they hold.
The buyers we talk to expect it to stay open for 9 to 18 months, so sooner is better.
Most companies have never had their data valued. A sister company of ABURA values company data independently, to international valuation standards, so its owner can license it, use it as collateral, or both.
You don't sell your house to raise money. You borrow against it, because it has been valued and insured. Company data can now work the same way.
A surveyor sets its value.
An independent valuer values it to international standards.
An insurer covers it.
An insurance wrapper backs that value.
The bank takes it as security for a mortgage.
A lender takes it as security for a loan.
You keep living in it.
You keep using it, and you can still license it with the lender's agreement.
No. It is a loan against an asset you own, so there is no equity to give away.
No. You borrow against it, the way you borrow against a home instead of selling it.
Often, yes. Many owners license through ABURA and borrow against the same data, with the lender's agreement.
Our sister company, independently of ABURA, to international valuation standards.
To AI labs and the data platforms that supply them, on terms the owner agrees.
Once valued and insured, company data can be taken by a lender as collateral, the way a bank lends against a home.
A friend, client or old colleague mentions code, drawings or operational data sitting unused, often at a company that has closed.
A short guided check, one question at a time. It tells you whether the lead is worth a conversation. It never gives a price.
Once they are happy to be introduced, the check writes the message for you. Send it to whoever invited you.
We speak with the owner in confidence, confirm what they own and introduce it to the right buyers. A sample set is needed only at the point of close, and the owner decides at every step.
Share a company's name only once the owner is happy to be introduced. Until then, the check keeps the name off the message for you.
No files, logins or screenshots. The data stays with its owner from the first conversation to the last.
Value is settled later, with the owner and the buyer. A number given early becomes the number everyone remembers.
If the code or drawings were built for a client, the client may own them. For a closed company, ask who holds the rights now. It is the question that matters most.
Health records, biometrics, card details and children's data are never part of it. Everything else is anonymised before it moves.
Very often, yes. A closed company's code, drawings and records usually still exist with the founders, a liquidator or in a backup, and whoever holds the rights can still license them.
Yes. A sister company of ABURA values company data independently, so it can be licensed or used as collateral, much as a bank lends against a valued, insured house. Owners can start a preliminary valuation on the valuation page, which opens without registration.
No. Rough answers are fine, and almost every question in the check has a "not sure" option. We work out the detail with the owner.
Never. It stays with its owner. You pass on a short message, and that is all.
Nothing at the start. A sample set is needed only at the point of close, and the owner decides what goes into it.
AI labs and the data platforms that supply them. We introduce owners to them in confidence, and only with the owner's agreement.
Something like: "AI labs are looking for data that has never been published, and what your company built may be exactly that. Would you be open to a short, confidential conversation to see if it qualifies? Nothing is shared without your agreement." The check writes a version for each lead, matched to what they hold.
The buyers we talk to expect it to stay open for 9 to 18 months, so sooner is better.
Most companies have never had their data valued. A sister company of ABURA values company data independently, to international valuation standards, so its owner can license it, use it as collateral, or both.
A bank lends against a house because it has been valued and insured. Company data can work the same way.
A surveyor sets its value.
An insurer covers it.
A bank takes it as security for a mortgage.
Our sister company values it to international standards.
An insurance wrapper backs the valuation.
A lender can then take the data as security.
To AI labs and the data platforms that supply them. The owner keeps the data and agrees every term.
Raise finance against data the company already holds.
Independent by design. The valuation is carried out by our sister company, separately from ABURA, and is never priced off a sale. The owner decides what, if anything, is shared between the two.
Fifteen questions, about ten minutes. Rough answers are fine, and anything you would rather discuss on a call can be skipped.