Seventeen separate artificial-intelligence researchers were sent at this question, each reading the actual laws, treaties, tax rules and government processing tools, and none of them told what answer to reach. Four of them found errors in their own first conclusions and corrected themselves. No lawyer has seen any of this and nothing here should be filed on. It exists so the two of you can have one conversation, and then buy an hour of real legal time knowing exactly what to ask.
What this answers
Munim is a British citizen living in Cleveland, Ohio. He is there on a student work permission that runs from the end of his Stanford MBA: a card valid until 15 June 2028, tied to working for Eulia, the company he and Janika own together. Nothing has ever been filed for him with the American immigration service. There is no application pending. The clock people assume is running has not started.
Janika is a Finnish citizen living in Espoo, working for Skanska. She also holds permanent residence in Britain, granted under the scheme for Europeans who were living there before Brexit and confirmed from her Home Office letter in May. She loses it only after five straight years outside the country, which is unusually generous, and she is nowhere near that.
What the two of you have said you want: to be living together within about a year, to start a family within one to two years, and for Janika to be able to work wherever you both are. That last one is a hard condition rather than a preference, and it is the thing that rules most of the options out.
Answer this before anything else
Are you willing to live together outside America for the year or two the American process takes, or do you need to be together in America?
Only the two of you can answer it. It costs nothing, it takes one conversation, and every expensive decision below is downstream of it.
If the answer is yes, then the three dates that have been driving all this anxiety, 15 September, 18 September and August 2028, stop mattering at once. All three only bite if you are relying on Munim's student status inside America.
If the answer is no, you are choosing a route where the best realistic case has Janika able to live and work in America in early 2028, and the likely case is later than that.
Almost everything confusing about the timelines comes from two misunderstandings. Clear them up and the rest of this page reads easily.
The petition asks a single question: does this person qualify? It is about Munim alone and nobody else. Paying an extra government fee forces a decision in forty five working days, and the typical case is decided in about 1.2 months. Crucially, having a petition on file puts no restriction on travel at all.
The application is the separate filing that turns a qualifying person into a permanent resident. Typical time: 5.8 months. This is the one that carries the work permit, the travel document, the medical examination and, sometimes, an interview. It is also the one that traps you in the country while it is pending.
People describe the two as though they were one thing, which is why the numbers you have been quoted never added up.
The waiting times published on the government's own website are not averages and not typical cases. In their own words, each figure is "the amount of time it took us to complete 80% of adjudicated cases". In other words, four cases in five finished inside that time and one in five took longer, and how much longer is not published anywhere. It is the slow end of normal, not the typical case.
The real typical times come from a different page, which says plainly that its figures are medians. For the year to 30 June 2026:
| What is being decided | Typical time | Note |
|---|---|---|
| The petition, with the fast-track fee paid | 1.2 months | The fee is worth it. Without it, the same petition takes about 29 months. |
| The green card application | 5.8 months | Cannot be fast-tracked at any price. |
| A work permit behind that application | 5.1 months | Cannot be fast-tracked. The published 80 per cent figure for this one is 12.5 months. |
| A travel document | 6.0 months | Which is why it cannot reach a December wedding. |
| The paperwork that lets a spouse follow later | 7.1 months | Before the embassy even opens her file. |
Plan on these. Budget for the published figures, which are longer.
For most people the agonising wait is the queue: America caps green cards per year and per country of birth, and Indian and Chinese applicants wait decades. The monthly table that shows how far down that queue the government has reached currently shows no queue at all for British applicants in the categories that matter. So this entire stage, the one that dominates everybody else's story, costs you nothing.
One caution. The September table carries an explicit warning that queues may come back later this year as demand shifts, and an approved petition holds your place if they do. But a refused one holds nothing, which is the correction set out further down and the reason the advice here is file as soon as the case is good enough to be approved, rather than file as early as possible.
Press the buttons to switch. Each one changes the charts, the dates, and everything written underneath them.
You stay in Cleveland, file everything here, and marry in Ohio in December. This is the fastest route to your own green card and by far the slowest route to the two of you living together.
Nobody asks a single question and every stage runs at the typical speed. Around half of these petitions attract a request for more evidence, so read this as the top of the range rather than as the plan.
One request for more evidence on the petition. It adds three to six months and it restarts the fast-track clock from zero.
Your green card application dies with the petition, because it has nothing left to stand on. Your student work permission survives, but only because you never left the country while it was pending. You start again from nothing, a year later.
Time runs left to right, from September 2026 to mid 2029. The three upright lines on every chart are 15 September 2026, then 15 June 2028 when your work card expires, then 14 August 2028 when your permission to be in America ends. Bars that reach the right-hand edge carry on past it.
You are in Cleveland throughout. Janika is in Espoo, visiting for up to ninety days at a time. You are not living together at any point until she is admitted as a permanent resident, which on this route is the last thing that happens rather than the first.
Only from the day she lands as a permanent resident, and there is no earlier moment. She cannot get a green card from inside America at all, because everyone who enters on the visa waiver is barred from applying here, and the only exception is marriage to an American citizen. So she has to be processed at an embassy after you become a resident, and the paperwork stage that starts that off has a typical wait of just over seven months before the embassy even opens her file.
Everything. Once the green card application is filed, leaving America counts as abandoning it. A wedding abroad, a family emergency or a funeral becomes a choice between going and losing the case.
Your own green card, sooner than on any other route here, and your American position is never put at risk by a move abroad.
A single request for more evidence on the petition. That one event is the difference between Janika arriving in early 2028 and arriving in late 2028.
You both move to Britain, where you can each live and work from the day you arrive, and the green card is processed at the American embassy in London while you get on with your lives.
Petition approved first time, embassy stage at ordinary speed. Notice that neither of you stops working at any point on any of these three charts.
One request for more evidence, plus ordinary embassy scheduling. The American date moves by eight months and nothing else changes at all.
This is the whole argument for this route. A refusal costs you the American plan and nothing else, because you are already living together and both earning. You try again on another route from a stable base.
Time runs left to right, from September 2026 to mid 2029. The three upright lines on every chart are 15 September 2026, then 15 June 2028 when your work card expires, then 14 August 2028 when your permission to be in America ends. Bars that reach the right-hand edge carry on past it.
Britain, together, from around December 2026. You are a citizen. Janika already holds permanent British residence, confirmed from her Home Office letter in May. Neither of you applies for anything, pays anything, or waits for anything to be granted.
Immediately in Britain, and on the day she lands in America. Because she is processed at an embassy rather than from inside the country, she arrives as a permanent resident rather than as somebody waiting on a work permit.
Nothing. There is no application pending inside America to abandon, so you can travel freely the whole time.
The bad branch stops being a disaster. If the petition fails you have lost a filing fee and some months, not a year of her working life.
Your American student work permission lapses once you stop using it, so if the green card never comes, that door has closed behind you. Eulia's operations also move abroad, which raises company tax questions that need separate advice.
You both move to Finland. The green card runs on the same clock as the Britain route, but Munim needs a Finnish residence permit before he can live there, and he cannot get one until you are married.
Assumes the Finnish spouse permit is granted quickly after a December wedding. Nobody has checked how long Finland actually takes on that permit, and that gap is the weakest number on this whole page.
One request for more evidence, plus the Finnish permit taking as long as these things usually take.
You are together and she is working, so a refusal is survivable in the same way it is on the Britain route. The difference is that Munim is the one on a permit rather than at home, and his own treaty route back to America has probably switched off.
Time runs left to right, from September 2026 to mid 2029. The three upright lines on every chart are 15 September 2026, then 15 June 2028 when your work card expires, then 14 August 2028 when your permission to be in America ends. Bars that reach the right-hand edge carry on past it.
Finland, together, but only once you are married and Munim's permit is granted. She is at home. He is the one who has to apply for something, and this is the only route on the page where his own right to be in the country can be refused by anybody.
Continuously. She never stops, which is the one thing this route shares with Britain.
Nothing, for the same reason as Britain: there is no American application pending to abandon.
Finland is the best of the five countries for having a child, by a wide margin. Childcare is capped near three hundred and thirty five euros a month, there are three hundred and twenty days of paid leave at roughly seventy per cent of salary, and it has the best newborn survival rate of the five.
Three things. Munim's Finnish permit is an application rather than a right, and it can be refused. Living in Finland probably switches off his British treaty route back to America, which is a real option he would be giving up, and the section below explains why. And Finland may already treat Eulia as a Finnish company because Janika directs it from Espoo, which is a live tax exposure quite separate from any of this.
The correction that takes the pressure off
You do not need the green card in your hand by 2028. You need the application filed and pending.
A properly filed application keeps you in lawful stay and carries its own work permit, independent of your student status. So the requirement is to file while you still have status, not to finish. Almost all the urgency you have been feeling comes from getting this one thing the wrong way round.
Working backwards from each of the three things you might want:
| What you want | Last date you could safely file | The step that breaks it first |
|---|---|---|
| Green card in hand before your work card expires, 15 June 2028 | About December 2026 comfortably. April 2027 at a push. | A request for more evidence on the petition. That alone adds three to six months and restarts the fast-track clock. |
| Green card in hand before your permission to be in America ends, 14 August 2028 | About February 2027 comfortably. June 2027 at a push. | The same. |
| Merely filed and pending, with a work permit issued | About January 2028, and legally right up to 14 August 2028. | The work permit, at five months, and it cannot be fast-tracked at any price. |
Are any of these already out of reach? No. None of them. Nothing has expired and nothing is about to. The third has well over a year of slack in it.
You asked for this directly and it has not been on any previous version of this page. These are one researcher's estimates, reasoned from your record against the current approval statistics, and the current statistics are far worse than the numbers most people have in their heads.
| Route | Estimate | What it rests on |
|---|---|---|
| The national-interest green card, which is the one this page is mostly about | 30 to 40% as things stand 45 to 60% after four to six months of work |
From the government's own data files: 45.5% approved in the first half of this government year, which is 5,410 approvals against 6,486 refusals. More than half are now refused. Last full year was 55.2%, and it fell through the year to a low of 35.7%. For comparison the rate was about 80% in 2023. |
| The extraordinary-ability green card | 12 to 22% | Approvals ran at 44.7% so far this government year, so it is no longer the safer alternative it used to be. The route asks whether you are at the very top of one single field. The honest answer is that you are unusually good at several things rather than the very best at one, and this route does not reward that. |
| The O-1 work visa, which is not a green card but keeps you in America | 60 to 75% | Much the same evidence, judged far more forgivingly, and a full application was professionally prepared and paid for in 2025 and never filed. No reliable published approval statistics exist for it, so this figure is reasoning rather than data. |
The two per cent figures in bold are the government's own published counts, read out of its performance data files by two separate researchers today who agreed. The estimates in the left column are one researcher's judgement about your specific record, and it reasoned against a slightly more favourable anchor than the 45.5% above, so treat its bands as marginally optimistic. Published rates for this category differ by more than thirty percentage points between law firms quoting the same period, so a number from a firm's website is not evidence.
Read as outcomes rather than as routes, by the time your work card expires in June 2028:
The gap between the first line and the third is the entire strategy. The green card probably does not finish by then. Staying in America and working almost certainly does, and it rests mostly on the O-1 rather than on the green card petition.
You asked exactly this. The honest answer has three parts.
The part of the test most applications fail is whether you are well placed to actually do the thing you say you will do, and yours is unusually good: a medical degree, a Stanford MBA, a master's in health economics, two Royal College fellowships, and a first-author systematic review cited in a Cochrane Review, which is independent evidence that your work influenced a field rather than a claim that it did. There is also no requirement anywhere in this route to be at the top of your field, which is the exact hurdle that kills the extraordinary-ability route for you. And the argument for waiving the normal employer sponsorship is unusually clean, because you own half the company, and asking an employer to advertise your job to American workers is close to incoherent when the employer is you.
The first part of the test asks whether what you propose to do matters nationally, and this is where it is difficult. The government's own manual says that general assertions about entrepreneurs creating jobs are not enough, and that a petition involving a startup "should demonstrate in detail how this specific endeavor meets the national interest requirement". Eulia's stated business is consumer wellness software, non-medical, that does not diagnose or treat. That description was chosen for tax reasons and has nothing to do with immigration, and it sits close to the manual's own examples of what fails.
The manual then lists six kinds of evidence it wants from a founder: ownership and role, degrees and licences, investment, accelerator participation, government awards or grants, and intellectual property. Three of those six are simply empty. Eulia is three months old, pre-revenue, has no outside investors at all, holds no patents, and has no government grants. The manual specifically treats outside investment as independent validation of an endeavour's importance, and there is none to point at.
This is where the 30 to 40 per cent becomes 45 to 60 per cent. The most valuable thing available is free, and it is not growth at all: it is closing the gap between the work your record actually supports, which is health-system efficiency and health economics, and the endeavour as currently described, which is consumer wellness software. Beyond that, the things that would move the number are the three empty evidence categories: an outside investor, an accelerator place, a government grant, or filed intellectual property. Users and revenue would help too, since there are none today to evidence traction.
One warning that has to travel with that advice. Changing how Eulia describes itself is a tax decision before it is an immigration one. The consumer wellness framing was built deliberately to support a large American tax relief on the eventual sale of the shares. The petition, the training plan Eulia signed for your student work permission, and the company's stated business must all tell the same true story, or the inconsistency becomes a liability rather than an advantage. This needs the tax and immigration questions answered together, by a person, before a word of it changes.
You asked whether a refusal now would weigh against you later. Nothing in the earlier research answered it, so a separate counsel researcher was sent at it today, with no view supplied and no hint as to what anyone hoped the answer would be. The answer is better than you probably feared, and it arrives with one correction that changes the advice you were about to be given.
A refusal does not bar you from trying again. There is no limit on how many times, and no waiting period. Nothing in the statute or the regulations caps it. In fact the law says the government shall approve a petition whenever eligibility is shown, so a previous refusal cannot lawfully be held against a later petition that does establish eligibility. The one one-strike rule in that part of the law is about marriage fraud and has nothing to do with you.
There is also no doctrine that binds a later officer to an earlier officer's refusal. The rules require a fresh decision on the new record. Appeal decisions in these cases bind only the parties to that case and do not set policy.
The hinge, and it decides how bad a refusal is
Keep your student status alive in its own right, and a refusal leaves you roughly where you started. Live on the green card work permit instead, and you are out of status the day the refusal issues.
A refusal on its own, while you are lawfully in student status, triggers nothing at all: no removal notice, no unlawful presence, no consequence beyond the wasted fee. The government's own policy on when it starts removal proceedings turns on whether the person is lawfully present when the unfavourable decision lands.
This overturns something you were told earlier today
A refused petition banks nothing. Only an approved one holds your place in the queue.
The regulation is one sentence long: "A denied petition will not establish a priority date." The popular reason for filing early is to bank your place before queues come back. That reason is simply wrong if the petition is refused, because there is nothing to bank.
So the right rule is not "file as early as possible". It is "file as soon as the case is good enough to be approved". An approved petition genuinely does protect you against queues returning. A weak early one protects nothing, costs the fee, and puts a permanent description of your company into a federal file. This page said the opposite two sections ago, before this research came in, and this is the corrected version.
There are three ways a filed petition can end badly and they are not equal. Best to worst:
Two mechanics attach to that. Deadlines for answering a request for evidence are twelve weeks, and thirty days for a notice of intended refusal, and no extension is available for either. And withdrawing is not a shield if something in the file was untrue: the fraud policy explicitly reaches cases that ended in a withdrawal.
Filing the green card petition cannot be used to refuse your O-1 work visa. The regulation says so directly: filing a preference petition "shall not be a basis for denying" an O-1, and someone may lawfully come for a temporary period and at the same time seek to become a permanent resident. That matters, because the O-1 is the route carrying most of the roughly 85 per cent chance of your still being lawfully in America and working in June 2028. Trying for the green card does not put that at risk. The same protection exists for the H-1B, with a statute behind it as well.
The form question, and this is the one genuinely permanent risk in the whole analysis. The moment any immigrant petition is filed for you, the truthful answer to the visa-application question asking whether one ever has been becomes yes, and it stays yes whether the petition is later approved, refused, withdrawn or abandoned. Filing is what triggers it, not the outcome. Answering yes costs you extra scrutiny and nothing more. Answering no, even carelessly, is a lifetime bar, and the test does not require any intention to deceive or that anyone believed you. Read the question off the live form when the time comes rather than off anyone's summary, including this one.
The researcher separated what a refusal costs automatically from what is merely likely from what is folklore. Two of the folklore items are worth naming because both circulate widely: "one refusal and you can never file again", which is false, and "at least I bank my place in the queue", which is false in exactly the case where you would want it to be true.
One caveat the researcher raised about its own work: the government's electronic regulations site was unreachable, so the regulation text came from the certified printed edition current to 1 January 2025. It saw no sign of an amendment to the sections it relied on, but that is a residual risk. It also found that published approval rates for this category vary by more than thirty percentage points between different law firms quoting the same period, and that at least one firm repeats the priority-date claim corrected above, so it used the government's own data files throughout and recommends discounting firms' figures entirely.
You asked this directly, and it was never answered before. A separate researcher was sent at it today with no view supplied, and it read the government's own approval spreadsheets and real firms' published price pages rather than working from impressions.
The government charges the same whoever prepares the papers. Filing online for the whole path, which is the petition, the green card application, the work permit and the travel document together, comes to $3,195. Adding the fast-track fee on the petition takes it to $6,160.
One detail worth knowing because it is widely got wrong, including on two law firms' own websites read today: the Asylum Program Fee that rides along with the petition is $300 for a self-petitioner, not the $600 a company would pay. The government's own fee schedule says so in terms. A firm's website is not a reliable source for a government fee.
Same form, same officers, first half of this government year
Backed by an employer's labour certification: 96.6% approved. Asking for the national interest waiver instead: 45.5% approved.
That is 26,744 approvals against 936 refusals in the first case, and 5,410 against 6,486 in the second. It is the identical form on the identical desks. The entire difference is the waiver argument, which is the part that says your work matters enough nationally that the usual employer sponsorship should be set aside.
And the market prices exactly that. One firm charges $4,000 to prepare the employer-backed version of that form and $10,000 to prepare the waiver version. The $6,000 gap between them is the price of the argument.
There is no published evidence, anywhere, that professionally prepared petitions of this kind are approved more often than self-filed ones. The government's own datasets carry no field recording whether an applicant had a lawyer, so the comparison cannot be made from them. Any firm quoting you a percentage improvement is quoting something untraceable. The often-repeated figures showing a large gap come from a study of immigration court cases, which is an adversarial hearing with a government lawyer on the other side, and they do not transfer to a paper decision made at a service centre.
So the decision has to rest on something other than a claimed success rate.
The forms themselves are close to worthless as a service; they are fillable and you would manage them. Assembly is worth little. The load-bearing thing is defining and framing the proposed endeavour, which is the exact thing the 45.5 per cent figure is measuring. Second is choosing which evidence goes in and how the recommendation letters are built.
The researcher read nine of the appeal decisions published last autumn in full. Seven of the nine died on the same point: national importance. The recurring mistakes are specific and worth knowing whichever route you take:
Two traps that apply to a founder specifically, and both are new information here:
| Route | All in | What you are buying |
|---|---|---|
| Prepare it yourself | $3,195, or $6,160 with fast-track | Government fees only. You write the argument the 45.5 per cent is measuring. |
| Yourself, plus a paid consultation first | Add free to $550 | One hour asking one question: state my proposed endeavour in a sentence, is it nationally important? This is the cheapest thing on the page and it tests the exact thing that fails. |
| Yourself, with a lawyer reviewing your draft | Add about $2,000 to $4,000 | A second pair of eyes on the framing before it becomes a permanent record. No firm publishes this price, so it is an estimate. |
| Yourself, with a lawyer held in reserve for a query | $550 to review, then $350 an hour | Really published, at that price. It makes the honest expected cost of self-filing about $3,195 plus roughly $4,000 if a query arrives, which is still under the cheapest full representation. |
| Full representation | Add $5,500 to about $15,000 | Real published flat fees today run from $5,500 (with a refund if refused, queries included) through $6,000, $10,000, $12,500 to $15,000. |
The single term to pin down in writing, whichever firm you use: is responding to a request for more evidence included? The published spread on that one clause is $0 to $5,000. Some firms include it. One bills it separately at $350 an hour, capped at three to five thousand.
A lawyer matters least when a case is obviously strong and least when it is hopeless. It matters most in the middle, and on a 45.5 per cent approval rate the middle is where most cases now sit.
But the researcher added a second consideration that is not about probability at all, and it is the one that should weigh most here. A founder's petition necessarily describes his own company in a permanent federal file. Whatever it says about what Eulia does and why it matters nationally is a record you cannot later retract, and it has to agree with the training plan Eulia already signed for your student work permission, and with how the company describes itself for tax. That is not a probability question. It is a question about writing something down permanently in a domain where the words are load-bearing.
So the cheap move, whichever way you eventually go, is the consultation. Free to $550, one question, and it tests the exact thing that kills seven cases in nine.
This has been stated confusingly before, so here it is plainly. There are two moments in the process, and only the second one gates the wedding.
The State Department manual, in a passage revised on 18 August 2026, twelve days ago, says a spouse acquired "before the principal applicant's admission to the United States or adjustment of status" gets the derivative status, and one acquired "after the admission/adjustment" does not.
On your dates, December is comfortable and October is not required. If everything were filed on 15 September, the petition alone takes about two months and the green card application then takes about six more. For your case to complete before a mid-December wedding, that second stage would have to be decided in roughly four weeks, which has never happened in the published data. The researchers found no scenario at all in which an October wedding is necessary.
Two further points from that same passage, both in your favour. There is no deadline by which Janika must follow; she could do it years later. But she must do it before Munim becomes an American citizen, which is five years away and worth remembering.
There are three separate traps here rather than one, and they only bite on the route where you file the green card application inside America.
The clean answer, if you are staying in America: marry in Ohio in December. Same day licence, no waiting period, no notice, no residence requirement because Munim already lives in Cuyahoga County, and both of you simply have to turn up at the probate court. Janika flies in as she would anyway. That single choice dissolves all three traps at once.
And if you are going abroad: simply do not file the green card application before you go. There is then nothing pending to abandon. The cost is the friendlier financial-means test, and that is the entire trade.
| England and Wales | Finland | Ohio | |
|---|---|---|---|
| Notice before the ceremony | 29 days | A check for legal impediments comes first | None |
| Residence needed beforehand | 7 days each, in the district | None stated | None. Munim already lives there. |
| How long the paperwork takes | Included in the 29 days | 5 to 7 weeks where one of you is foreign | Same day |
| Home Office referral? | No. Because Janika holds settled status, you are both exempt, so no referral and no 70 day extension. | Not applicable | Not applicable |
| Working back from mid-December | In England by about 1 November, notice about 8 November | Start in early October | Turn up |
Ohio is by a wide margin the fastest and cheapest, and it is the only one of the three that requires no travel by Munim. Britain is entirely workable and the seven days of residence each is the only real constraint.
These have been run together in every previous version, including by us. They are three different things and only one of them is a genuine decision.
15 September 2026
A new rule ends the open-ended student admission everyone has had until now. Its transition clause says that anyone properly maintaining their status on 15 September keeps it until the later of their work card expiring or their course end date, plus sixty days. For Munim that works out at 15 June 2028 plus sixty days, which is 14 August 2028. There is nothing to file and no form to send. It happens automatically.
So the thing to protect is the phrase "properly maintaining status", and the weak points there are not travel. They are the known defects in the student work arrangement: a salary of twelve thousand dollars against the forty hours a week the signed training plan certifies, which works out at about five dollars seventy seven an hour and engages a rule requiring pay comparable to an American worker doing the same job; an employment eligibility form that is not finished; and a check that was never run. Those are what to fix before the 15th.
Travel is a smaller matter than earlier versions of this page suggested. An approved work card readmits you to the card's own end date. The cost of one trip is that your grace period at the end drops from sixty days to thirty. The researchers disagreed about this, one of them reading the published wording as pointing at a course end date already in Munim's past, so it is a question for the lawyer rather than a settled fact.
17 September 2026
This is the last day a green card application gets judged under the current, friendlier rules on financial means. It applies only if you file inside America. If you are processing at an embassy abroad, this date does not touch you at all.
18 September 2026
From this date a rule takes effect that scraps the current framework: the closed list of benefits that count, the minimum factors an officer must weigh, the requirement to write down reasons for a refusal, and the list of people who are exempt. What replaces it is broad officer discretion, under a law whose mandatory considerations include "assets, resources, and financial status". A salary below the federal poverty line for a household of one is exactly the profile that discretion could bite. Filing by the 17th locks in the older framework.
This has been asserted before without being explained. There are three separate reasons and they are independent of each other.
Munim is a citizen. Janika already holds permanent residence. Neither of you applies for anything, pays anything, waits for anything, or can be refused anything. Of the five countries considered, this is true nowhere else. In Finland Munim needs a permit he cannot get until you are married. In Switzerland the same. In Australia Janika would need a partner visa costing over eleven thousand Australian dollars with a typical wait of seventeen months. In America Janika cannot work at all for about thirteen months.
This is the fact that reframes everything and it is easy to miss. Janika can never get a green card from inside America. Anyone whose last entry was on the visa waiver is barred from applying from within the country, and the single exception is being married to an American citizen, which Munim can never be for her. So processing at an embassy is not a fallback or a compromise. It is the only mechanism that exists for her. And it is the better one anyway: she arrives as a permanent resident able to work on the day she lands, rather than arriving and then waiting a year for a work permit.
This is the finding nobody expected. There is a class of American visa for people from countries that have an investment treaty with the United States. Britain is on that list, but its entry carries a footnote that no other country's does. The treaty, which dates from 1815, applies only to "British territory in Europe" and to "inhabitants" of that territory, and the State Department defines an inhabitant as "one who resides actually and permanently in a given place, and has his domicile there".
Read that twice, because the consequence is unusual. Munim is a British citizen everywhere, but he is only a British treaty national while he actually lives in Britain. Living in Britain switches this route on. Living in Switzerland or Australia very likely switches it off, and living in Finland probably does too, because a British citizen domiciled in Helsinki is not an inhabitant of British territory in Europe. He cannot borrow Janika's Finnish nationality either, because the rules require the company and the person to share one.
Janika's own position is the mirror image and quietly better: Finland is on the same list with no footnote at all, so her treaty nationality travels with her wherever she lives.
Be clear about what this does and does not give you today. It keeps a door open; it does not open it. That treaty visa is for someone who has invested a substantial sum of their own money in an American business, and going through Eulia it currently fails on three separate grounds that have nothing to do with where you live: neither of you holds the majority the rules require, the investment has to be substantial and at risk and nobody has established that the money exists, and the business cannot be marginal, meaning it has to do more than support the two of you. That last one is the binding problem and paperwork cannot cure it, because a company that cannot pay Janika a salary is the definition of what that rule excludes. So the footnote matters as a thing you keep rather than a thing you can use now.
One honest caveat. The wording of that footnote is certain and quoted above. How strictly the negative half is enforced against a British citizen living elsewhere is not: no case applying it that way was found. It is the single most valuable thing to ask a real lawyer, because it is what most of this section turns on.
There is no route back to America you can simply choose. You would have to build one, and it takes about a year: Eulia opens a genuine British arm, you both actually run it for twelve months, and then the American company brings you over as staff transferring within the same business. That twelve months cannot be shortened or bought. Moving to Britain starts that clock rather than stopping it, which is the point.
| Country | Can you both live and work there? | Company and personal tax | Having a child |
|---|---|---|---|
| Britain | Both of you, today, with nothing to file. | $8,050 a year now 47.5% if profitable |
Free maternity care from day one, no ceiling, intensive care included. Neither of you is the isolated one. |
| Finland | Janika yes. Munim no, unless you marry. The living-together route needs two years of it, which you have never had. | $13,750 a year now 51.3% if profitable |
Best of the five. Childcare capped near 335 euros a month, 320 days of leave at about 70% of salary, best newborn survival. The first three years may even be cash positive. |
| United States | Munim yes. Janika no, in any way that permits work, for roughly thirteen months. | $3,950 a year now 30.7% if profitable |
Worst of the five by a distance. A birth costs your whole $7,300 deductible. No paid leave, and Eulia is too small to owe even unpaid leave. |
| Australia | Munim yes, on the residence he already holds. Janika no: a partner visa is AUD 11,710 and takes about seventeen months. | $10,100 a year now 37 to 43% if profitable |
A two year trap. Time spent outside Australia does not count, so neither of you could claim parental pay for a baby born within two years of arriving. That is about AUD 26,122 simply not there. |
| Switzerland | Janika effectively yes. Munim no since 2021, unless you marry. | $19,450 a year now 37.0% if profitable |
Safe and expensive. There is also a yearly tax on what you own, Eulia shares included, payable in cash whether or not the company pays you anything. |
The first tax figure is the annual cost of running the company from that country with no profit. The second is the total rate on a modelled four hundred thousand dollars of profit with one hundred thousand dollar salaries, a profit that does not exist today. Exchange rates are assumed and some 2026 bands are estimated. Illustrative, not advice.
A Delaware company is an American taxpayer permanently, because of where it is registered. The 21 per cent is a floor that nothing on that list gets under. Moving to Switzerland does not buy a low rate; it buys 21 per cent plus a second tax system, a second set of accounts and a second audit surface. And in your actual situation, which is no profit, moving would cost four to fifteen thousand dollars a year to produce exactly the same amount of tax as now, which is none.
Since 2021 Finland can claim a foreign company if its board directs it from Finland. Janika has been a director in Espoo since May. The American treaty with Finland is the harshest of the four here: if this bites, the company gets no treaty protection at all. Two separate tax researchers found this independently. The fix is about where decisions are formally made, and Finland gives binding answers in advance, which is the cheap way to settle it.
If Eulia were managed from Australia it would become, under the treaty, a resident of nowhere: no tie-breaker, no dispute route, up to 46 per cent combined. The mirror image is a clean line you control, which is to keep the decision-making out of Australia. Munim already holds residence there, so unlike every other country there is no visa moment to mark the transition.
| Situation | You pay |
|---|---|
| Normal birth | $7,300 |
| Caesarean | $7,300 |
| A $117,878 intensive care stay for the baby | $7,300 |
| A pregnancy crossing 1 January, so two deductibles | $9,000 to $10,000 |
| If the family deductible turns out to be the combined kind | About $14,600 |
Against a stated salary of $12,000 and about $10,700 a year of adult premiums. Deaths in childbirth per hundred thousand births: United States 17, Britain and Finland 8, Switzerland 5, Australia 2.
And no, there is no subsidy sitting unclaimed. An earlier version of this page said there was, then said the reason was your immigration status. Both were wrong and independent tax counsel has now settled it. For 2026 you fail on income, not status: a rule that used to let a lawfully present non-citizen below the poverty line claim the help was repealed with effect from this year, and $12,000 is about 77 per cent of the $15,650 poverty line. From 1 January 2027 you fail on category instead, absolutely. Paying the full $447.25 is the correct outcome, a pay rise buys nothing, and there is no phone call to make. Separately, there are a few things you are entitled to and are not using, which are being worked through elsewhere; one of them has to be cleared against the 18 September change above before anyone touches it.
You asked me not to give a view until all the evidence had arrived. It has, including two pieces commissioned today that were missing when the last version of this page was written. Here is the view, and then the reasoning, and then the case against it.
My recommendation
Marry in December. Live in Britain. Run the green card through the London embassy, and file the petition when it is good enough rather than by any date on a calendar.
This is not a compromise between America and Britain and it is not a retreat from America. On the objective you actually stated, it is better than the America-first plan on every axis I can find, and it gives up less than it looks like it gives up.
Every route into America that is fast enough to matter puts Janika in Cleveland, legally forbidden from any paid work, for about thirteen months. No income, no colleagues, no social security number, not driving, and no family nearby but your brother. That is also, on your own stated plan, the year you would be trying to have a child.
In Finland the stranded one would be Munim, waiting on a permit. That is a far more survivable version of the same problem, because she would be the one recovering from childbirth and she would be at home. In Britain nobody is stranded at all. Of everything in these seventeen research passes, this is the fact I would trade away last.
More than half of these petitions are now refused. So the honest way to compare the routes is not to compare their good branches, which look similar, but to ask what a refusal costs you.
On the America route a refusal in mid-2027 costs you the green card application as well, leaves you with a student permission running out in June 2028, and leaves Janika exactly where she started, having waited a year for nothing. On the Britain route the same refusal costs you a filing fee and some months, while you are already living together and both earning. Same probability, wildly different consequence.
You are a citizen. She holds permanent residence. There is no application, no fee, no waiting period and no possibility of refusal. Every other arrangement on this page depends on somebody saying yes: an American officer, a Finnish one, an Australian one. This is the only one that does not.
She cannot get a green card from inside America under any circumstances, because of the visa waiver bar and because the one exception requires marriage to an American citizen. So the route this recommendation takes is the route she has to take anyway. Choosing it costs nothing that was ever available.
A policy issued in May tells officers that granting a green card from inside the country is a matter of discretion and grace, and that even a visa category which permits an intention to settle permanently is not by itself enough to earn a favourable exercise of it. A student visa does not permit that intention at all. That argument did not exist in earlier versions of this page and it points the same way as everything above.
Living in Britain switches on a treaty route back to America that living in Switzerland, Australia or probably Finland switches off. It is explained three sections up, and it is the one place on this page where the country you live in changes what you are entitled to ask for. It happens to favour the same answer.
Somebody advocating the America route would say this, and it is not a weak argument:
The friendlier financial-means test expires in eighteen days. There is no queue right now. You can file the petition and the green card application on the same day. Munim is already here with valid work permission running to June 2028. Delay is irrational when the door is open and visibly closing.
My answer is that it optimises the wrong thing. It buys Munim a faster American green card at the price of Janika being unable to live and work beside him throughout, which is the one thing you both said you actually want. It also freezes him inside the country for a year or more, during which a wedding abroad, a family emergency or a funeral becomes a choice between attending and losing the case. You would be buying a document at the cost of the life the document is for.
And there is a second answer that only arrived today: the eighteen-day deadline is real but it is about the quality of the outcome rather than whether it is possible at all, and rushing a weak petition to catch it now looks worse than it did this morning, because a refused petition holds no place in any queue.
These are the honest weak points, including three places where the researchers disagreed with each other.