I-90 Final-Stage Wait Rises 13 Days, Yet Sits 60% Below Last Year

The I-90 (Application to Replace Permanent Resident Card) final-stage wait rose from 69 days to 82 days in the week of September 14, a 13-day gain that is within the range of moves this series has seen over the past 12 weeks. One year ago this same week, the median was 211 days, and the 3-year same-week average is also 211 days. Applicants in 2026 are waiting roughly 60% less than those who were in queue at this time last year. The week's one-week move is real, but the year-over-year picture is the more meaningful frame for most readers tracking their own cases.

The 12-week series for I-90 has been anything but smooth. Medians ranged from 15 days in the week of June 29 to 104 days in the week of August 17, with this week's 82 landing near the upper end of that band. The 8-week regression slope is +3.8 days per week, but R² sits at just 0.25, meaning the data does not fit a clean upward line. The pattern is labeled 'noisy stable': week-to-week readings move substantially without establishing a confident direction. The 25th percentile (p25) this week is 77 days and the 75th percentile (p75) is 116 days, a spread that widened compared to recent tighter weeks.

The gap between p75 and the median is 34 days this week (116 versus 82), whereas in the week of August 31 the p25 and p75 were clustered tightly at 65 and 70 days respectively. This week's wider interquartile range (the spread between the 25th and 75th percentiles) reflects genuine dispersion in outcomes rather than a uniform shift in processing speed. The mid-August reading of 104 days demonstrated that the series can produce single-week outliers that then recede, and the current reading may follow the same pattern.

An RFE (Request for Evidence) is USCIS's standard follow-up letter asking for missing documents; receiving one does not signal denial. This week, cases that had an RFE response in their file posted a 22-day median final-stage wait, versus 47 days for cases that never received an RFE. That 2.09x speed ratio in favor of RFE-stamped cases is the anomaly the system flagged as rfe_fast. The most plausible mechanical possibility is that a batch of RFE responses submitted in an earlier period was adjudicated together this week, compressing that group's apparent wait. The data brief does not tell us the cause, and a single week is not enough to confirm any mechanism.

RFE-touched cases cleared in 22 days this week, faster than the 47-day median for plain cases: a statistical curiosity of the week's mix, not evidence of a policy shift.