FY2026 Year End: Why a Bigger FY2027 Pool May Not Move India EB-2
FY2026 ends on September 30. The August inventory report is out, USCIS has published ten months of processing data, and DOS has released monthly issuance through February. That is enough to set down what the year has done so far, rather than what it was expected to do.
Four things stand out. Most of the pending queue has a priority date that is already current. Only a small share of it gets decided in the final two months. Whether the last of this year's numbers get used comes down to a September surge that has not happened yet. And the family numbers going unused this year set up a materially larger employment pool for FY2027, though not in the way most people assume.
Sources: USCIS Pending Applications for Employment-Based Preference Categories (August 5, 2026); USCIS Report to Congress on Service-Wide Forms (through July 2026); USCIS Employment-Based Approved I-140 Petitions by Country report, FY2026 Q2; DOS Monthly Immigrant Visa Issuance Statistics (October 2025 through February 2026); DOS Annual Numerical Limits FY2026; INA 201(c), 201(d), 202(a)(5)(A), 203(b). Not legal advice.
1. What is pending, and how much of it is already eligible for a visa number
| Category | Pending | Date is current | Date not yet current |
|---|---|---|---|
| EB-1 | 59,307 | 43,140 | 16,167 |
| EB-2 | 76,791 | 46,981 | 29,810 |
| EB-3 | 55,965 | 35,888 | 20,077 |
| EB-3 Other Workers | 7,535 | 5,149 | 2,386 |
| EB-5 (all) | 22,879 | 21,989 | 890 |
| Total | 222,477 | 153,147 | 69,330 |
About 69 percent of the pending employment queue has a priority date that is already current. Those people are waiting on USCIS to decide the case, not on the visa bulletin to reach them. The concentration matters: Rest of World holds 106,124 of those 153,147 cases. India holds 16,225 and China 20,630.
One clarification, because the official wording invites a misreading. USCIS marks these cases as having a visa "available", but that means the priority date is current, not that a visa number is set aside for them. Numbers are consumed when a case is approved, and the annual limit still applies. So there are 153,147 cases eligible to receive a number, not 153,147 numbers waiting to be used. Section 3 works through what that difference means for the rest of this year.
The extremes inside that. Rest of World EB-1 (30,778) and EB-2 (41,966) are 100 percent current, because those categories carry no cutoff at all. India EB-2 is zero of 27,981, which is what an Unavailable category looks like from the inside.
2. What gets decided between August 6 and September 30
The inventory above is a snapshot taken on August 5. The fiscal year ends on September 30. That leaves a window of 55 days, or about 1.8 months, and everything below follows from it.
Step one, how many numbers FY2026 has already used. Adjustment approvals ran to 101,707 across the nine reported months from October through July, and one month in that stretch has not been published, which puts the ten month figure near 113,000. Consular issuance in the employment categories was 17,070 across the five months DOS has released, and the five unpublished months add somewhere between 15,000 and 17,000 more.
| FY2026 through July 31 | Visa numbers |
|---|---|
| Adjustment of status approvals (USCIS) | about 113,000 |
| Consular issuance (DOS) | 32,000 to 34,000 |
| Used so far | 145,000 to 147,000 |
| FY2026 worldwide limit | 186,317 |
| Still available for the final 55 days | 39,000 to 41,000 |
Step two, how many of those can physically be issued. This is where the pace matters, and USCIS has shown two very different paces this year. Its FY2026 average is 11,301 adjustment approvals a month. Its best month, March 2026, was 20,885.
| Aug 6 to Sep 30 (1.8 months) | At the FY2026 average | At the March 2026 pace |
|---|---|---|
| Adjustment approvals | 20,400 | 37,800 |
| Consular issuance | 5,400 | 5,400 |
| Total issued in the window | 25,900 | 43,200 |
So against 39,000 to 41,000 numbers still on the table, the year closes somewhere between using about 26,000 of them and using every one. Section 3 takes up which end is more likely.
Step three, how much of that goes to Rest of World. Rest of World holds 106,124 of the 153,147 cases with current priority dates, which is 69.3 percent. Adjudication capacity is national and shared, so on a proportional split Rest of World collects roughly that share of the decisions.
| Rest of World, date already current | On Aug 5 | Decided by Sep 30 | Rolls into FY2027 |
|---|---|---|---|
| EB-2 | 41,966 | 5,600 to 10,400 | 39,000 to 42,000 |
| EB-1 | 30,778 | 4,100 to 7,600 | 23,200 to 26,700 |
| EB-3 | 25,516 | 3,400 to 6,300 | 19,200 to 22,100 |
| EB-3 Other Workers | 4,120 | 550 to 1,000 | 3,100 to 3,600 |
| EB-5 (all) | 3,744 | 500 to 900 | 2,800 to 3,200 |
| Total | 106,124 | 14,200 to 26,200 | 87,300 to 97,600 |
The EB-2 row does not simply subtract decisions from the August count, and the others do. That is the correction a reader earned: Rest of World EB-2 is current, so people keep filing into it, and its pending count grew by 6,616 between April and August even while approvals ran at their heaviest. A subtraction from a snapshot understates what a current category carries forward. The same effect applies to the EB-1 and EB-5 rows, which are current too, so treat those as the conservative end.
Step four, the whole queue on October 1. Across every country, 222,477 applications were pending on August 5. Between 20,400 and 37,800 of them get an adjustment decision in the window, which leaves roughly 185,000 to 202,000 still pending when FY2027 opens, before counting anyone who files in the meantime.
The cases roll into the new year. The visa numbers do not. That distinction is the whole of section 3.
3. Will the numbers actually get used?
Section 2 leaves 39,000 to 41,000 numbers on the table and a window that can absorb anywhere from 25,900 of them to all of them. Which end the year lands on is genuinely open, so here is the evidence both ways.
The case for the numbers being used. The 11,301 monthly average understates what USCIS can do. In March 2026 it approved 20,885 employment I-485s in a single month, nearly double its own average. Two months at that pace is about 41,800, which would absorb the entire remaining pool. There is also no shortage of cases to work on, since 153,147 have current priority dates. And last year gives a hint about behaviour: September 2025 approvals fell to 2,662 immediately after 11,827 in August. That is the signature of a category running out of numbers, not of an agency running out of capacity, which suggests the system pushes until the supply is gone.
The case for some going unused. The window at the recent average rather than the March peak issues about 25,900 against 39,000 to 41,000 available, which would leave something in the region of 13,000 numbers unissued. The consular half is the weaker link, because issuance there has been running below last year and immigrant visa interviews are being rescheduled worldwide.
So the base case is that most of the pool gets used, and the risk is concentrated in the consular channel rather than in adjustment of status. We are not putting a confident number on the shortfall, because it depends on a September surge that has not happened yet and that we cannot observe until USCIS publishes the month.
This is also where the two headline figures reconcile. There are 153,147 cases with current priority dates and at most 41,000 numbers left. Eligibility is a queue; the limit is a flow. The cases that do not get a number are not losing something they held, they carry into FY2027 and draw on a fresh pool.
Whatever does go unused does not carry into next year's employment limit. Under INA 201(c) it falls into next year's family limit. Employment numbers never bank for employment.
The traffic runs the other way too, and that is the channel that matters for FY2027. Under INA 201(d), family numbers unused this year are added to next year's employment limit. Family consular issuance has already dropped sharply. Immigrant visa applicants also reported receiving interview rescheduling notices worldwide from around August 24, and the State Department has said it is adjusting appointments around consular officer training sessions. No duration or scope has been published, so we have not assigned it a number, but the direction is lower issuance in the closing weeks.
| Family immigrant visas issued | Month |
|---|---|
| October 2025 | 16,424 |
| November 2025 | 14,487 |
| December 2025 | 15,766 |
| January 2026 | 12,551 |
| February 2026 | 7,720 |
February is the first full month under the January 21 processing restrictions, and it ran at roughly half the October to December pace. Projecting the rest of the year from those months, the family-to-employment fall-up into FY2027 lands somewhere between about 40,000 (restrictions lift and issuance returns to the autumn pace) and about 98,000 (issuance stays near February's level). Our working figure of about 73,000 sits inside that range. We are publishing it as a range because the seven months that would settle it have not been released.
So FY2027 opens with 140,000 by statute plus roughly 40,000 to 102,000, for a pool somewhere near 180,000 to 242,000, against 186,317 this year.
4. Where the FY2027 dates could land
We are not going to pick one pool and print one date off it. The fall-up is the single biggest unknown in this piece, and the seven months of DOS data that would settle it have not been released. So every figure below is computed at all three ends of our own range: a low pool of 180,131, a base of 213,066, and a high of 242,383. Read the column that matches what you think happens.
Correction, September 3 2026. This section previously assumed a single 90,000 fall-up and a 230,000 pool, and printed one set of dates off it. Two things were wrong. 90,000 was the middle of a range we had already revised: our working figure is 73,066, and the old table was sitting at the optimistic end without saying so. Rest of World EB-2's rollover was also understated at 31,600 to 36,400, when its pending count in fact grew by 6,616 between April and August while the category was current. Both are fixed below, and the arithmetic now runs out of our estimator's own code rather than by hand, including the rule that a projection may not print below a level no observed year has ever produced. This is the last revision to this section before the FY2027 limits are published.
Step one, split by category. Under INA 203(b), EB-1, EB-2 and EB-3 each take 28.6 percent of the pool. That is 51,517 at the low end, 60,937 at base, 69,322 at the high end. EB-4 and EB-5 take 7.1 percent each.
Step two, the per country floor. No country may take more than 7 percent, which inside a category works out to 3,606, 4,266 or 4,853 across the three pools. India, China, Mexico and the Philippines are each guaranteed that much, which leaves 37,093 to 49,912 for everyone else combined.
One honest simplification in that step, raised by a reader. Mexico and the Philippines will not use their full floor each. They hold about 2,500 pending EB-2 cases between them against roughly 8,500 of floor between them at the base pool, so about 6,000 numbers go unused, and by statute unused numbers go to the oversubscribed countries in priority date order. The mechanism is real. What the record shows is that it has not moved India. That slack existed in all six years we can measure, and in the four years Rest of World used its full share India still finished between 0.86 and 1.21 times its own floor: FY2019 slack 2,676 with India at 1.04, FY2020 slack 3,264 at 0.86, FY2023 slack 5,229 at 1.08, FY2024 slack 4,518 at 1.21. The two years India ran far past its floor, FY2021 at 6.75 and FY2022 at 7.77, are the two years Rest of World could not use its own allocation. Rest of World absorption is the variable that moves India, not the Mexico and Philippines remainder.
Step three, work out what Rest of World actually wants. If Rest of World does not use its share, the leftover does not sit idle. Under INA 202(a)(5)(A) it goes to the oversubscribed countries, in priority date order, with no 7 percent cap. So the question for India and China is not how big the pool is. It is whether Rest of World can absorb its own share.
Rest of World demand has three parts, and all three are measurable. The I-485 cases rolling over from section 2. The approved I-140 holders who have not filed an I-485 yet, from the USCIS quarterly report. And consular applicants, who never appear in the I-485 inventory at all.
That middle part is the one that decides this, and it is worth looking at on its own:
| Rest of World, approved I-140 with no I-485 filed | FY2025 Q2 | FY2026 Q2 |
|---|---|---|
| EB-1 | 0 | 0 |
| EB-2 | 28,456 | 6,734 |
| EB-3 | 21,767 | 21,558 |
Rest of World EB-1 has nobody waiting with an approved petition, and Rest of World EB-2 has fallen from 28,456 to 6,734 in a year. That is not a mystery. Both categories became current, so the people holding approved petitions filed, which is the same event that pushed the Rest of World EB-2 pending count up to 41,966. A current category drains its own waiting room. Rest of World EB-3 has barely moved, 21,767 to 21,558 over the same year, because EB-3 still has a cutoff and its waiting room has had no reason to empty.
Step four, add it up. Principal petitioners bring family members, so a petition is worth more than one visa number.
| Rest of World FY2027 demand | EB-1 | EB-2 | EB-3 |
|---|---|---|---|
| I-485 cases rolling over | 23,200 to 26,700 | 39,000 to 42,000 | 24,400 to 27,300 |
| From the I-140 waiting room | 0 | about 8,500 | about 29,900 |
| Consular | 4,052 | 11,556 | 10,627 |
| Total wanted | 27,300 to 30,800 | 59,100 to 62,100 | 64,900 to 67,800 |
| Its share of the category | 37,100 to 49,900 | 37,100 to 49,900 | 37,100 to 49,900 |
| Spills to India and China | 6,300 to 22,700 | none | none |
Step five, turn that into dates. India and China split whatever spills in proportion to how many of their cases are waiting on a number, which today is 74 percent to India in EB-1, 86 percent in EB-2 and 81 percent in EB-3. Each country's pending I-485s are then walked in priority date order until the visas run out. Where nothing spills, a country does not simply collect its bare floor, because no observed year has ever worked out that way; the range there runs from the lowest to the median multiple of its own floor that the country has actually received since FY2019.
| Category | Date now | Low pool 180,131 | Base pool 213,066 | High pool 242,383 |
|---|---|---|---|---|
| India EB-1 | Oct 15 2022 | Apr to May 2023 | Jul to Sep 2023 | Nov 2023 to Current |
| India EB-2 | Unavailable | Oct to Nov 2013 | Nov 2013 | Nov to Dec 2013 |
| India EB-3 | Jan 1 2014 | Mar to Apr 2014 | Apr to May 2014 | May to Jun 2014 |
| China EB-1 | Jul 1 2023 | Sep to Oct 2023 | Current | Current |
| China EB-2 | Sep 1 2021 | Nov to Dec 2021 | Current | Current |
| China EB-3 | Jan 1 2022 | Current | Current | Current |
What this says, and it is not what the headline number suggests
India EB-1 is the one that moves, and the reason is an empty waiting room. Rest of World EB-1 wants 27,300 to 30,800 against a share that runs from 37,100 to 49,900, because it has no approved petitions left to convert. Everything above that spills, and India takes 74 percent of it. That is 13,981 to 16,572 numbers for India EB-1 at the base pool, which walks the queue to about July to September 2023. It takes the high pool to reach November 2023, and only the very top of it clears the 19,261 filed queue outright.
India EB-2 stays near its floor. EB-3 does a little better. Rest of World EB-2 wants 59,100 to 62,100 against that same 37,100 to 49,900, so nothing spills at any pool we can justify, and India collects 4,266 to 4,605 at base. That is roughly two months of priority date from where it stopped at September 2013. EB-3 is the same story with a smaller queue behind it, 5,189 to 6,185 numbers and April to May 2014.
One rule is doing real work in those EB-2 and EB-3 rows. When nothing spills, the arithmetic wants to hand India its bare 7 percent floor, and no observed year has ever come out that way: India EB-3 has received between 1.22 and 4.23 times its own floor in every year since FY2019, EB-2 between 0.86 and 7.77. An earlier version of this section ignored that and printed India EB-3 at March 2014, which a reader was right to push back on. The floor guard is now in the code the table is generated from, not applied by hand afterwards.
For scale on that: India EB-2's last printed final action date before it went Unavailable was September 1 2013, and it had reached July 15 2014 in April and May of 2026 before retrogressing. A November 2013 landing is a modest advance on where it stopped, and still short of where it briefly stood earlier this year.
What the State Department has already said, and how it fits. Section F of the July 2026 Visa Bulletin states that India's pro-rated EB-2 limit was reached, that the category is unavailable for the remainder of FY2026, and that it is likely that in October the final action date will advance to at least the date announced in the May 2026 bulletin, which was July 15 2014. That is further forward than the November 2013 above, and the two are not in conflict, because they answer different questions. The State Department is saying where the October date is likely to print. Our figure is how far a full year of numbers carries the queue. A final action date can print ahead of what the year's supply will actually deliver and then retrogress, which is exactly what happened this year: India EB-2 stood at July 15 2014 in May and was Unavailable by July. The same bulletin conditions its own statement on demand and on the FY2027 annual limit, neither of which is published yet.
So a bigger pool is worth a great deal to EB-1 and very little to EB-2. The pool size is not what separates them. What separates them is that Rest of World already emptied its EB-1 waiting room and has not emptied its EB-3 one.
Does this method reproduce reality?
It is worth checking the arithmetic against what actually happened rather than trusting it. Realized number use for India in recent years:
| India | FY2022 | FY2023 | FY2024 | Model above |
|---|---|---|---|---|
| EB-1 | 21,437 | 16,604 | 8,809 | 13,981 to 16,572 |
| EB-2 | 59,431 | 4,301 | 3,916 | 4,266 to 4,605 |
| EB-3 | 12,571 | 4,765 | 3,643 | 5,189 to 6,185 |
EB-2 and EB-3 land close to the FY2023 and FY2024 actuals. EB-1 lands between the FY2022 windfall and the FY2024 low, which is what an above normal pool with an empty Rest of World waiting room should produce.
FY2022 is the cautionary column. India EB-2 took 59,431 that year, and no arithmetic about pool size explains it. That happened because Rest of World could not use its numbers during the consular shutdown, so almost the whole category spilled. Today Rest of World has 106,124 cases with current dates and is filing steadily, so that condition is absent. A large pool is not the same thing as a large spillover.
What would change these numbers
The largest one now runs in India's favour. Counting the whole Rest of World I-140 waiting room as FY2027 demand is conservative, because that queue converts over more than one year rather than all at once. If only part of the Rest of World EB-3 waiting room files next year, Rest of World EB-3 stops being oversubscribed and some of it would flow to India EB-3, moving that date past the April to May 2014 above. EB-2 has less room to move that way, because its waiting room is nearly empty already and the rollover alone almost fills its share.
Why FY2028 may matter more than FY2027 for India EB-2
The analysis above points somewhere the headline numbers do not. Rest of World EB-2 wants more than its own share in FY2027 almost entirely because of what it carries forward: 39,000 to 42,000 cases rolling over from this year, on top of its consular channel and a nearly empty petition backlog.
That rollover is a one time load. If Rest of World works through it during FY2027, then in FY2028 its EB-2 demand stops being driven by a stock of cases that are already filed and starts being driven by how fast new petitions are approved. At that point Rest of World could fall short of its share of the category, and under INA 202(a)(5)(A) the shortfall passes to the oversubscribed countries with no per country cap.
Correction, September 3 2026. This paragraph originally said the waiting room behind that rollover holds only 6,734 approved petitions. A reader was right to push on it. The 6,734 is measured at March 31 2026 in USCIS Report C, before Rest of World EB-2 became current, so most of those people filed during the window that follows and are already counted inside the 41,966 pending rather than waiting behind it. We do not have a post April measurement of that waiting room, and the sentence should not have carried a number. The direction of the FY2028 point is unchanged; the figure supporting it was not usable.
This is not a new mechanism. It is the one that produced FY2021 and FY2022, when Rest of World EB-2 used 23,919 against 31,000 to 43,650 in the surrounding years, and India EB-2 received 28,246 and then 59,431.
Three things would all have to hold: Rest of World clears most of its rollover during FY2027, its petition backlog stays thin, and family issuance stays suppressed enough to keep the employment pool large. None is guaranteed, and this is a mechanism to watch rather than a forecast. But if you are in India EB-2 asking which year is worth watching, the arithmetic points at FY2028 rather than FY2027.
Limits that apply throughout. Queue sizes here are the full USCIS inventory for each country and category, which includes the workbook's "Prior Years" bucket of priority dates older than its oldest year column. An earlier version of this article omitted that bucket and understated India EB-1's queue by 1,299 cases, enough to turn "reaches late 2023" into "goes current". The date walk counts filed I-485s only, so consular applicants and petition holders who have not filed sit on top of these counts and would slow every date shown. And India's filed queue stops at January 2015, the current filing cutoff, so no method can project past that point from filed cases alone. This walk is an illustration built from the statute and the published inventory so the steps can be checked. It is not how the Estimator on the site computes its own figures, and the two can differ. It is scenario arithmetic at one assumed fall-up, not a forecast.
What to do with this
If your priority date is already current, the bulletin is not your constraint. Processing capacity is, and 79,900 to 92,000 eligible Rest of World cases carry into FY2027 alongside you. Being current is also not permanent. A category can retrogress, and several did during FY2026.
If you are in any China category, FY2027 is a year to watch closely: those queues are small enough that the base pool plausibly finishes all three. India EB-1 is close behind, reaching roughly mid to late 2023 at base and finishing only at the top of the range.
If you are in India EB-2, the honest expectation is a few months of date movement rather than years, even on a strong pool, because Rest of World wants more than its own share and little spills onward. India EB-3 should do somewhat better, on the order of four to eight months of date movement, because Rest of World EB-3 has left numbers unused in every year we can measure. Neither is a year of movement, and neither is nothing.
You can run your own category and priority date through the Green Card Estimator. Estimates are approximate and built from the published government data cited above.
This article is informational and is not legal advice. Immigration outcomes depend on individual circumstances. Consult a licensed immigration attorney about your case.