I got endorsed for what was then the Tier 1 (Exceptional Talent) visa back in 2016, landed in London that May, and ten years on I'm still here building things. The visa worked exactly as advertised. It gave me the freedom to change companies, found things, take roles in fintech, and do the kind of public technical work that keeps compounding without ever having to think about immigration friction.

The programme has been renamed, almost shut down, contracted out, and rebranded since I got mine. The endorsing body has nearly disappeared and come back. The form has changed. The evidence rules have got stricter. But the underlying bar, the thing they are actually trying to assess, is roughly the same as it was a decade ago. The route is still one of the best things the UK does for technical people, and the application is still mostly about presenting your evidence clearly rather than gaming the system.

This is what's changed since 2016, what they actually want to see in 2026, and what I'd put in the application if I were doing it today.

What's changed since I got it

In February 2020 the Tier 1 (Exceptional Talent) visa was renamed to the Global Talent visa, and Tech Nation took over from Tech City UK as the digital technology endorsing body. The criteria got tweaked but stayed broadly the same shape, with the same Talent and Promise pathways and the same idea of mandatory plus optional criteria.

Then in January 2023 Tech Nation announced it was shutting down because the government had pulled its funding, which caused a fair amount of panic in the tech and immigration community. Founders Forum acquired Tech Nation later that year and kept it operational. In May 2025 Tech Nation won the new £11m contract to continue as the digital technology endorsing body for another three years, which finally settled the question of who runs this thing. In August 2025 the separate Tech Nation application form was scrapped and everything moved to a single GOV.UK Stage 1 form, which is genuinely simpler than what came before.

The other big change is the evidence rules tightened up significantly in early 2025. All evidence now has to be from the last five years, anything that looks like it was created specifically for the application gets penalised, and there is now an explicit prohibition on using AI or language processing tools in your application, which they will reject for. Rejection rates are reported to be rising as a result.

Roughly 4,000 people apply through the digital technology route each year. Endorsement success rates for digital tech sit around 54 to 65 percent depending on the source, with academic and research routes much higher and arts and culture roughly in the middle. The visa stage itself is mostly a formality once you have endorsement, so the real hurdle is Tech Nation, not the Home Office.

The two pathways and which one to pick

There are two pathways, Exceptional Talent and Exceptional Promise, and you have to choose one on the application form.

Exceptional Talent is for established people with at least five years of substantive work in digital technology. You should be able to point to a track record of recognised contributions over that period.

Exceptional Promise is for people earlier in their careers, typically with less than five years of experience, who have already started doing recognised work and have evidence of where they are heading.

The visa rights are identical on both pathways. The only difference is how quickly you can apply for indefinite leave to remain, three years for Talent and five for Promise, which matters but isn't the main consideration. You can be downgraded from Talent to Promise by the reviewers if they decide your case is stronger as Promise, but you cannot upgrade later, so once you're endorsed under Promise you stay on the five year ILR clock.

In practice, if you've got a decade of solid work behind you with public, verifiable evidence, apply under Talent. If you're earlier on but already shipping things that are getting recognised, Promise is the right call. Don't reach for Talent without the track record because they will either refuse or downgrade you, and a downgrade is fine but a refusal costs you the application fee and a lot of preparation time.

The criteria, plainly

You need to meet one mandatory criterion and at least two of four optional criteria, with at least two pieces of evidence supporting each one, up to a total of ten evidence pieces.

The mandatory criterion is essentially that you are recognised internationally as a leading talent (Talent) or a potential leading talent (Promise) in digital technology, with evidence from the last five years. This is the headline claim, and it has to be evidenced rather than asserted.

The four optional criteria, pick two:

1. A proven track record of innovation as a founder, senior executive, or employee at a product-led digital technology company, or as someone working in a new digital field or concept.
2. Proof of recognition for work outside your immediate occupation that contributes to the advancement of the field, which is where open source, public writing, talks, and mentoring sit.
3. Significant technical, commercial, or entrepreneurial contributions to a product-led digital technology company.
4. Academic contributions to the field through research that has been endorsed by an expert.

You also need a CV, a personal statement of intent, and three letters of recommendation from senior people in the industry. The letters cannot all come from the same company, and they need to speak to specific contributions rather than offering generic praise.

The five year recency rule is strictly enforced, so anything older than that effectively doesn't count for the headline criteria, even if it's significant. This catches people who have done great work earlier in their career and assume it will carry the application.

What they actually want to see

This is where most applications fall down, so it's worth spending time on. The thing to internalise is that Tech Nation reviewers are reading hundreds of applications, working through a structured rubric, ticking boxes against criteria. They are not inferring significance from context, and they are not generously interpreting your achievements. If a piece of evidence doesn't clearly map to a criterion and clearly demonstrate it, it adds nothing.

If I were applying today, the shape of my evidence pack would be roughly this. Companies I've founded, with public products, real users, and verifiable revenue or traction. Open source projects on GitHub with stars, forks, and external contributors, where I can point to specific things the work has enabled in the broader ecosystem. The years I spent at Visa Innovation in London, where the public appearances, panel work, and external engagement during that period are independently documented. The technical work at BVNK on payment processing infrastructure, with letters from senior people there describing specific contributions. The Emergent Ventures grants, which are independently awarded and verifiable. Sustained public writing on ksred.com over many years, with traffic figures and reach. Open source bioinformatics work like EabhaSeq which is independently funded and demonstrates contribution to a new field.

Each of those points at a specific criterion, each is independently verifiable, each is recent, and each can be evidenced rather than just asserted.

The presentation of the pack matters as much as what's in it. This is the thing I keep telling people. The strongest applications are structured like a barrister's bundle rather than a CV. Each evidence piece is its own appendix, clearly labelled, dated, with a short cover note explaining which criterion it supports and why. The personal statement is the argument, the CV is the timeline, and the evidence pieces are the exhibits that prove the claims. Heavy appendices are not a sign of weakness, they are the entire point. You are not writing a marketing document about yourself, you are assembling a case file.

What they explicitly don't want, based on the updated 2025 guidance. Salary, equity, and bonuses are no longer accepted as proof of significant contribution because they don't actually prove impact. Online-only mentoring through platforms like ADPList no longer counts as recognition beyond your day job. Generic letters of recommendation that don't speak to specific contributions are weighted near zero. Anything that looks created for the application, including suspiciously recent talks at low-tier events or articles in places nobody reads, gets discounted. And as of the latest guidance, any sign that the application itself was written using AI tools will get you rejected outright, which is worth taking seriously given how easy it is to fall into that habit.

The mistake people make

The recurring pattern I've seen when giving informal advice is people treating this like a job application. They list achievements in bullet points, write a smooth personal statement, attach a CV, and assume the reviewer will infer significance from context. Reviewers don't infer. They check claims against evidence and move on.

The applications that work do something different. They make explicit arguments. They state a claim, point at a specific evidence piece in the pack, explain why that evidence meets the criterion. The personal statement reads like a structured case rather than a polished bio. The evidence pieces themselves are organised so a reviewer can verify each claim quickly without having to dig through context. People underestimate how much of the assessment is just whether the reviewer can map your claims to your evidence in the time they have.

The other recurring mistake is not having a clear UK plan. The application asks what you're going to do here, and vague answers about contributing to the ecosystem don't cut it. They want specific projects, specific people you've already spoken to, specific things you'll build or work on. If you can show that you've already started conversations with people in the UK tech sector and have concrete plans, that lands much better than aspirational statements about the value you'll add.

The mechanics

Since August 2025 it's a single GOV.UK Stage 1 form rather than a separate Tech Nation portal. The endorsement fee is £524, the visa fee is £766, and the Immigration Health Surcharge is currently £1,035 per year per person. Endorsement decisions take 5 to 8 weeks, the visa decision after that takes up to 3 weeks, and you have 3 months from receiving endorsement to apply for the visa itself.

If you're refused at endorsement stage you have 28 days to request a review, which is free and reportedly succeeds in roughly half of cases where it's used. You can also reapply from scratch after addressing the issues. There's no annual cap on endorsements anymore, so you're being assessed on the criteria rather than competing against other applicants for a fixed slot.

Once endorsed and on the visa you can apply for indefinite leave to remain after 3 years on Talent or 5 years on Promise, which is a significantly faster route to settlement than most other UK work visas.

Conclusion

Ten years in, the Global Talent visa has been one of the most valuable things I've ever done. It made the move to the UK possible, it gave me freedom to change companies and found things without sponsorship friction, and the route to settlement was clean. The bar is real, the evidence requirements have got stricter over the years, and the rejection rates are not trivial. But it isn't a mythical bar. If you've been doing public, technical, recognised work for five years and you have evidence to show it, you have a case worth making.

The application itself is mostly about presentation, structure, and clear mapping of evidence to criteria, not about being secretly more impressive than you actually are. The worst thing you can do is wait until your evidence feels "good enough" before starting, because the only way to find out whether you have a case is to actually assemble the pack and put it in front of them.

If you're thinking about applying, start by listing every piece of public, verifiable work you've done in the last five years, then map each one to a specific criterion. If the gaps in the matrix are obvious, you'll know what to work on. If the matrix is mostly full, you're probably ready.