Imminent Research Grants 2026: $100,000 Split Into Five $20,000 Awards for Language Technology Research, Open Worldwide and Closing 1 October 2026
Imminent, the research centre of the translation company Translated, allocates $100,000 a year to fund five original one-year projects in language services — one grant of $20,000 in each of five research areas — with a 250-word pitch as the first-stage application and a submission deadline of 1 October 2026.
Imminent Research Grants 2026: $100,000 Split Into Five $20,000 Awards for Language Technology Research, Open Worldwide and Closing 1 October 2026
Most research funding for language and translation technology arrives through national research councils, sits behind an institutional affiliation requirement, and takes six to nine months to decide. The Imminent Research Grants do almost none of that. Imminent — the research centre run by the translation company Translated — puts up $100,000 every year, splits it into five $20,000 grants, opens the call to anyone anywhere in the world including companies and unaffiliated individuals, and asks for a 250-word pitch as the first thing it reads.
The current call is open and the submission deadline is 1 October 2026.
That combination is unusual enough to be worth understanding properly. A $20,000 grant will not fund a research group, but it is real money for a focused one-year piece of work: a dataset, a benchmark, an evaluation framework, a working paper, a prototype platform. Judging by the projects Imminent has funded so far, that is exactly the scale of work it is looking for.
Key details
| Item | Detail |
|---|---|
| Programme | Imminent Research Grants |
| Funder | Imminent, the research centre of Translated |
| Total annual allocation | $100,000 |
| Number of grants | 5 |
| Value per grant | $20,000 |
| Project duration | 1 year |
| Distribution | One grant assigned per research area |
| Status | Open |
| Submission deadline | 1 October 2026 |
| Geographic eligibility | Anywhere in the world |
| First-stage application | Online form with a 250-word pitch |
| Shortlist size | 20 candidates advance to the second evaluation phase |
| Second stage | Detailed project description plus a 2-minute video |
| Possible third step | 10-minute interview, if the committee wants more information |
| Payment schedule | 50% on signing the grant agreement, 50% on receipt of the final report |
| Grantee obligations | Two milestone meetings (remote permitted) and one published article within a year |
| Contact | [email protected] |
What the money is for, and how it actually reaches you
Imminent describes the programme as funding “five original research projects that explore the most advanced frontiers in the world of language services.” The framing matters: this is a company-funded research programme run by a business whose stated mission is to let everyone understand and be understood, and the grants are aimed at problems that sit at the boundary between language research and language technology in practice.
The payment structure is worth planning around before you apply. Imminent transfers 50% of the funds after the grant agreement is signed, and the remaining 50% on receipt of the final report. That means $10,000 up front and $10,000 at the end, which is fine for a project whose costs are spread across the year, and awkward for a project that needs to buy something expensive in month one or pay annotators in a single burst. If your budget is heavily front-loaded, say so and plan for a bridge.
Two milestone meetings are scheduled during the project to follow up on progress. Imminent explicitly allows these to be held remotely, so there is no travel cost buried in the commitment. Grantees are also asked to publish an article about their project on the Imminent Research Center website within a year of the final presentation. Treat that as part of the deal rather than an afterthought — it is a dissemination requirement, and a reasonable one for a $20,000 grant.
What Imminent’s published call does not specify is whether overheads or institutional indirect costs can be charged against the $20,000, how the payment is handled for tax purposes in different countries, or what happens to intellectual property generated by the project. None of that is confirmed on the public page. If any of it materially affects whether you can accept the grant, email [email protected] before you invest time in the application rather than after a decision.
The five research areas, and what Imminent means by each
Exactly one grant is assigned per category, which changes the strategic picture. You are not competing against the entire applicant pool — you are competing against everyone else who filed under your subject. Choosing the right category is a genuine decision, not a formality.
Imminent publishes problem examples and an “opportunity” statement for each area:
Language economics. Problem examples include the economic benefit of multilingualism and new indicators and models for it. Alternative focuses named are investment, planning, discovery, human resources management, supply chain management, and international trade. The opportunity Imminent states is improving management and entrepreneurship in a multilingual environment. This is the category most open to economists, trade researchers, and policy analysts rather than computer scientists.
Language data. The stated problems are how to collect and organise language data across modalities, how to investigate the utility and properties of data with respect to machine learning algorithms, and whether the impact of a given or planned dataset on a learning algorithm’s performance can be determined in advance. Alternative focuses include generating high-quality data from plentiful low-quality data and optimising human annotation effort in utility and economic terms. The opportunity: teaching machines using examples of human communication.
Machine learning algorithms for translation. The problem framing is improving the efficiency of learning algorithms for machine translation across quantity, quality, speed, adaptability, and context awareness. This is the most conventionally technical of the five, and probably the most crowded.
Human-computer interaction. Problem examples centre on engaging professional human translators in a symbiosis with machine translation: equitable incentives for human participation in developing machine translation, rational and emotional rewards, interface design, and seamless human-computer interaction. Alternative focuses are design, engagement, empowerment, interface, interaction, and gamification. The opportunity statement is blunt — “give value back to humans.”
Neuroscience of language. How the brain works in language learning and translation, with alternative focuses in experiments, theories, applications, neuroscience, psychology, and social psychology. The stated opportunities are how humans come to trust translations, how they open up to new languages, and what differentiates humans from machines in language learning.
Imminent adds that applications can be made in any field of research, so long as the project addresses one of these subjects. A trade economist, a psycholinguist, and an ML engineer are all in scope — in different categories.
Who can apply
The call is open to “startuppers, researchers, innovators, authors, university labs, organizations, and companies,” and applicants may apply from anywhere in the world. There is no stated requirement for a university affiliation, no PhD requirement, no career-stage restriction, and no country restriction published on the call page.
That breadth is the programme’s most distinctive feature and its main practical implication: an independent researcher, a two-person startup, and a full university lab are all evaluated on the same form. Nothing in the published criteria awards points for institutional prestige. The 250-word pitch is doing almost all the work at the first stage, so an unaffiliated applicant with a sharp, concrete proposal is genuinely competitive.
One additional flexibility is stated explicitly: the project can either be part of an existing research project you are already working on or a standalone project focusing on one of the proposed topics. If you already have a funded programme running, you can carve out a discrete, deliverable-bearing piece of it and propose that.
How the selection process works
A committee of leading experts evaluates the proposals. The process runs in stages:
- First stage — the form. All applicants submit the short online form, with the 250-word pitch at its centre.
- Second stage — the shortlist of 20. Twenty candidates are selected and asked for a more detailed description of the project plus a two-minute video in which the researcher gives a short speech about their project.
- Possible third step — an interview. Candidates may be contacted for a ten-minute interview if the committee requires further information.
- Award. Five grants are assigned, one per research area.
Twenty shortlisted for five awards means roughly a one-in-four conversion once you have cleared the first cut, and the video is where a lot of that separation happens. Imminent does not publish the number of applications it receives, so the true first-stage odds are not confirmed — but the structure tells you where to spend effort. Get through the pitch, then take the video seriously.
Imminent does not publish a results announcement date for this cycle on the call page. Plan on a decision timeline you do not control, and do not build a start date into any other commitment.
Writing the application
The form asks three things: a 250-word pitch of your project, what makes you the right team to solve it, and what you will produce concretely.
Two hundred and fifty words is roughly a page double-spaced, or one long paragraph plus two short ones. There is no room for a literature review, a methods section, or a background paragraph about why language matters. Every sentence that is not doing load-bearing work should be deleted.
A structure that fits the three questions:
- One or two sentences on the specific problem, stated as a gap rather than a topic. “Existing NLI benchmarks are English-only and label-only” is a gap. “Natural language inference is important” is a topic.
- Two or three sentences on the approach. Name the method, the data, and the scale. Reviewers in this field can tell whether a plan is executable in a year on $20,000 within about ten seconds of reading the method.
- One or two sentences on why you. Prior work in the exact area beats general seniority. A link to a running prototype or a published dataset beats both.
- A concrete deliverables list. This is the question Imminent asks explicitly, and the funded projects answer it explicitly.
On that last point, look at how past winners phrase it. One 2026 language economics project committed to “an open-access working paper and dataset” plus “a policy brief for governments and AI developers.” A 2026 machine learning project named its deliverable — MT-AT, an adaptive-testing framework for evaluating machine translation with fewer samples. A 2025 language data project promised the first multilingual multi-annotator NLI benchmark with annotator-provided explanations. Every one of them names an artefact that will exist at the end of the year and that someone outside the project can use.
What past winners look like
The funded portfolio is a better guide to fit than any eligibility list, and Imminent publishes it in full on its projects-awarded page.
Recent language economics awards went to a quantitative framework for estimating the global welfare and productivity gains from expanding multilingual AI coverage into underrepresented languages, and to LinguaEconomy, a project building a Global Linguistic Efficiency Index for Trade. Neither is a machine learning project; both are data-driven economics with a public dataset at the end.
In neuroscience of language, an award went to Michelle Cohn and Georgia Zellou at UC Davis for an experimental study of how first- and second-language English speakers judge LLM translation accuracy, including whether text-to-speech voice quality shifts trust and whether speakers of low-resource languages rely more on surface cues. Another went to researchers at the University of Parma and Roma Tre studying the neural basis of literal and metaphorical language processing in a second language.
In machine learning for translation, Luca Benedetto at Télécom SudParis won for applying Item Response Theory and computerised adaptive testing from educational assessment to machine translation evaluation, cutting evaluation cost while preserving measurement precision.
In language data, Barbara Plank at LMU Munich won for the multilingual multi-annotator NLI benchmark. In human-computer interaction, Simran Khanuja at Carnegie Mellon won for a human-in-the-loop platform letting translators localise images with text instructions, built on the finding that the best AI models achieve only 5–10% accuracy at localising images across countries. And a 2025 language economics award went to the AFLANG Project, building Africa’s first unified multilingual dictionary platform with verified lexicographic data, native-speaker audio, and offline functionality.
The pattern across all of them: a narrow, testable question; an open artefact as the output; and a clear line to how translation or multilingual communication gets better as a result.
Common mistakes to avoid
Trusting the wrong page for the deadline. Imminent’s general “Imminent Research Grants” overview page has been showing a previous cycle’s deadline while the “Apply for your grants” page carries the current one. The apply page states the submission deadline is 1 October 2026. Use that page, and if the two ever disagree again, email [email protected] rather than guessing.
Picking the category by comfort rather than fit. Because one grant is awarded per area, a project that sits awkwardly in a crowded category may have been competitive in an adjacent one. An economics-flavoured study of annotation cost could plausibly be language data or language economics. Read both problem-example lists and choose deliberately.
Proposing a five-year agenda. The grant is $20,000 for one year with a mandatory final report. A pitch describing a research programme rather than a project signals that nothing will actually be finished.
Leaving the deliverable vague. “We will advance understanding of…” is not a deliverable. A dataset, a benchmark, a framework, a working paper, a platform release — those are.
Ignoring the video. It is not a formality at the shortlist stage; it is a second, independent read on whether you can explain your own work. Two minutes is short enough that it needs a script.
Assuming an affiliation is required or disqualifying. Neither is true. Companies and independent innovators are named in the eligibility text alongside university labs.
Frequently asked questions
When is the deadline? 1 October 2026, per Imminent’s application page.
How much is each grant? $20,000, for one year, from a total annual allocation of $100,000 across five grants.
Can I apply from outside Europe or the US? Yes. Imminent states you can apply from anywhere in the world.
Do I need a university position? No. The call names startup founders, innovators, authors, organisations, and companies alongside university labs.
Can I submit more than one proposal, or apply in more than one category? Imminent’s published call does not address this. Ask by email before submitting a second application.
Is the $20,000 paid up front? No. Half is transferred after the grant agreement is signed and half on receipt of the final report.
Can existing work be proposed? Yes. The project can be part of an existing research project you are already working on, or a standalone project on one of the topics.
What do I have to do if I win? Deliver the project, attend two milestone meetings (which can be remote), submit a final report, and publish an article about the project on the Imminent Research Center website within a year of the final presentation.
When are results announced? Not confirmed on the public call page.
Next steps before 1 October 2026
Start by choosing your category and reading Imminent’s problem examples for it word by word — those lines are the closest thing the programme publishes to assessment criteria. Then write the deliverables sentence first and build the pitch backwards from it, because that is the sentence a committee assigning one grant per area will use to separate proposals.
Draft the 250 words, cut them to 200, and see what you lost. If the pitch survives losing a fifth of its length, it was carrying filler; if it collapses, you had it right. Then, before you submit, sketch the two-minute video you would record if shortlisted. If you cannot say the project out loud in two minutes, the written pitch is probably still too abstract.
Send eligibility or administrative questions to [email protected] well before the deadline — the same address handles questions about both the grants and individual applications.
Official links
- Application page and current call details: imminent.translated.com/apply-for-your-grants
- Programme overview: imminent.translated.com/imminent-research-grants
- Previously funded projects, by year and research area: imminent.translated.com/imminent-research-grants-projects-awarded
- Questions: [email protected]
Always confirm the deadline and terms on Imminent’s own application page before submitting. Where this guide says a detail is not confirmed — results timing, overheads, intellectual property, multiple submissions — that reflects what Imminent publishes, not an omission you should fill in by assumption.