About the institution
An institution built around evidence.
Frontier Labs publishes independent research on the technologies that will shape the next several decades. We are not a consultancy, not an advocacy organisation and not a vendor. We publish what the evidence supports, with the evidence attached.
Origin
The research began in 2022. Frontier Labs is the roof that was built over it — a single institution for work that had until then been conducted, and published, in pieces.
The work started from a narrow observation: the most consequential technologies of this decade were being described almost entirely by the people building them and the people selling them. That is not a conspiracy. It is a structural consequence of where the information sits. The people with access to training runs, deployment telemetry, supply chain contracts and interconnection agreements are the people operating them. Everyone else is working from announcements.
The result is a public record that is confident where it should be uncertain, precise where the evidence is thin, and silent about the assumptions carrying the conclusions. When that record is used to allocate capital, to write regulation and to plan infrastructure with decade-long lead times, the cost of its errors is not academic.
Frontier Labs exists to publish a different kind of record. Assembled from primary sources. Stated with its uncertainty attached. Open to being checked by anyone who wants to check it. We would rather publish a narrow finding we can defend than a broad one we cannot.
For the first years that work ran as separate enquiries with separate outputs. Bringing it under one name changed what could be attempted: a single methodology applied consistently, one source register, one correction log, and the ability to hold a question open across editions rather than closing it to meet a deadline.
We are deliberately small, deliberately slow, and deliberately uninterested in being first. Most of what we study will still matter in a decade. That is the timescale we work on.
Mission
To publish independent, deeply researched analysis of the technologies reshaping economies, institutions and the future of intelligence.
Vision
A public record of technological change that a careful reader can trust without having to trust us.
Trust in an institution is not a substitute for verifiability; it is what accumulates when verifiability is offered consistently over a long period. Everything we publish is designed to be checked. The methodology is published, the sources are registered, the intervals are stated, and the conditions under which we would change our assessment are set out in advance.
Research principles
Six commitments, applied without exception.
Evidence before opinion
We do not publish a view we cannot support with material a reader could check. Where the evidence is thin, we say the evidence is thin. Where we hold a view that the evidence does not yet establish, we mark it as a judgement and separate it from the findings.
Independence
Frontier Labs does not accept funding contingent on findings, does not offer pre-publication review to subjects of research, and does not publish commissioned work under its research imprint. Our funding structure is disclosed and our conflicts register is public.
Transparent methodology
Every publication carries its methodology in full: sources, assumptions, intervals, and the analytical choices that could have gone another way. A reader who disagrees with our conclusion should be able to locate precisely where they disagree.
Long-term thinking
We choose questions by their durability rather than their currency. Most of what we study will still matter in a decade. We would rather be slow and right than fast and cited.
Neutrality
We describe rather than advocate. On contested questions we set out the strongest version of each position and identify the empirical claims on which the disagreement actually turns. We do not have a policy programme.
Correction as practice
We are wrong sometimes. When we are, we correct it with the same prominence as the original claim, we record it permanently, and we examine why the error occurred. Published work is never silently revised or withdrawn.
Editorial standards
Separation of finding and judgement
Findings are things the evidence establishes. Judgements are things we believe on the basis of incomplete evidence. Every publication separates the two explicitly, and judgements are labelled as such in the text rather than in a footnote.
Adversarial reading before publication
A designated reader who was not involved in the research argues the strongest case against every finding before it is accepted. Where that case cannot be dismissed, it is published alongside the finding.
No pre-publication review by subjects
Organisations that are the subject of research do not see the work before publication and are not offered the opportunity to comment on it privately. Factual corrections are welcomed publicly, after release, through the corrections process.
Stated falsification conditions
Every substantive assessment is published alongside a statement of what evidence would change it. This is a discipline against unfalsifiable positions as much as a service to the reader.
Permanent record
We do not remove or silently revise published work. Corrections are appended with the same prominence as the original claim and recorded permanently in the change log.
Coverage
Twelve domains under continuous study.
Independence and funding
Frontier Labs does not accept funding contingent on findings, does not offer pre-publication review to the subjects of research, and does not publish commissioned work under its research imprint. Where a member of the research group holds an interest relevant to a publication, that interest is disclosed in the publication itself and recorded in the conflicts register.
Editorial control over what is researched, what is concluded and what is published rests entirely with the Frontier Labs research group. No funder, partner or affiliate has ever seen a publication before release, and none has the ability to delay, alter or suppress one.
Work with us, or tell us we are wrong.
We answer research correspondence, factual corrections, data requests and press enquiries. Corrections are treated as the most important category of message we receive.