Bucharest-Ilfov took 33.1% of September’s new registrations, essentially unchanged from its summer run. But beneath a flat aggregate, Ilfov held its volume while Bucharest shed more than a fifth of its count, and Cluj slipped from second to fourth. The dispersion that matters this month is inside the top tier, not between capital and provinces.

The framing that usually carries this story — capital versus provinces, centre against periphery — does not survive contact with September’s numbers. What the county-level data actually shows is a top tier that is re-sorting itself internally while the aggregate capital share stays put.

A share that is conspicuously stable

Bucharest recorded 1,779 registrations in September, or 25.7% of the national total. Ilfov added 510, or 7.4%. Together the two account for 33.1% of everything registered in the country that month.

Compared with the preceding months of 2026, that is unremarkable. Bucharest’s monthly share ran 23.4% in May, 23.6% in June, 25.2% in July and 28.1% in August before settling at 25.7% in September. The August figure stands out only because the national total was unusually small — 8,103 registrations against September’s 6,912 — which mechanically inflates any single county’s percentage. Averaging the May-to-September window, Bucharest lands at roughly 25.2%, within a rounding error of September’s reading.

The same holds for Ilfov: monthly shares of 6.8%, 7.5%, 6.8%, 7.0% and 7.4% across May to September. Ilfov is the most stable large county in the dataset. There is no dispersion story in the capital’s aggregate weight.

The real movement is volume, and it splits the two capital counties

Where September departs from pattern is in raw counts, and here Bucharest and Ilfov diverge sharply.

National registrations stood at 6,912 in September, down 14.7% from August’s 8,103. Some of that is seasonal: September is historically a middling month, ranking ninth of twelve by mean registrations, so a summer-to-autumn step down is not itself a signal.

But Bucharest’s decline was far steeper than the national move. Its count dropped from 2,276 in August to 1,779 in September, a fall of 21.8% — about seven percentage points worse than the country as a whole. Ilfov fell from 567 to 510, a decline of 10.1%, meaningfully milder than the national rate.

That gap has a compositional explanation. Bucharest’s August total was inflated by an unusually large PFA cohort — 1,083 sole-trader registrations that month against 1,185 SRLs, an almost even split. In September the PFA count fell to 752 while SRLs fell less severely to 1,016. Ilfov, by contrast, is overwhelmingly an SRL county: 361 of its 510 September registrations were incorporated companies, against just 145 sole traders.

The entity-type contrast between the two capital counties is stark and persistent. Across May to September, sole traders made up between 45% and 48% of Bucharest’s registrations but only 23% to 28% of Ilfov’s. Ilfov’s business fabric is built on the incorporated form; Bucharest’s is split almost down the middle as the PFA share has climbed through 2026. September’s Bucharest decline was, in the main, a sole-trader decline.

Cluj falls out of the top tier

The county that genuinely deviated from its own recent pattern is Cluj, and the deviation is negative.

Cluj registered 330 companies in September, placing fourth behind Bucharest, Ilfov and Timiş. Across the three preceding months it had run 564 (June), 577 (July) and 383 (August). September’s count represents a 13.8% fall from August and a 42.8% fall from July.

Timiş, by comparison, is flat. It registered 337 in September, exactly the same count as in August, after 602 in July and 522 in June. Timiş thus overtook Cluj for second place among provinces in September, but on the strength of Cluj’s retreat rather than any Timiş advance. The two counties had run neck-and-neck in the 520–600 band through June and July; September separated them by just seven registrations.

Cluj’s September figure is not merely a seasonal dip. Its share of the national total fell to 4.8% in September from 5.2% in July and 5.0% in June, so the decline outpaced the national contraction. That is a deviation from its own pattern, not a reflection of the calendar.

Cluj also carries the heaviest exit burden of any large county relative to its size: 394 lifecycle events in September against 330 registrations, a churn ratio of 119.4. Bucharest’s ratio was 77.5 and Ilfov’s 72.0. Cluj is registering fewer companies in September than it is losing, and losing them faster than the capital region does.

What sector rotation tells us about the capital’s role

National sector data for September complicates the received idea that certain sectors belong exclusively to the capital. Transport and warehousing led all sectors nationally with 1,181 registrations, ahead of retail at 1,004, information and communications at 737, and professional, scientific and technical activities at 722. None of these is a capital monopoly; transport and retail are distributed across the country, and Timiş and Constanţa routinely register substantial transport volumes.

The one place where the capital’s grip looks structural is the qualitative, high-value end rather than the volume end: across the top ten sectors, the steepest year-on-year decline in September was in manufacturing, at -68.7%, ahead of other services at -64.5% and education at -59.8%, while information and communications fell least among the major categories, at -29.9%. The sectors holding up best are those most concentrated in Bucharest and Cluj. But since the whole national total is down 53.3% year on year, these are relative differences within a broad-based contraction, not evidence of the capital absorbing activity from elsewhere.

The picture in one line

Bucharest-Ilfov’s September share of 33.1% sits inside a May-September band of roughly 30% to 35%, and neither county moved more than a percentage point from its own summer average. The capital region’s aggregate weight is a fixed feature of the registration landscape.

The deviation that matters is Cluj’s. At 330 registrations, down 42.8% from July and carrying the highest churn ratio of any large county, Cluj has fallen from second to fourth among all counties, outpacing the national decline. Timiş’s unchanged 337 in consecutive months is the counter-example: stability, not dispersion.

The distinction worth drawing is therefore not between the capital and the provinces. It is between a capital region whose aggregate weight is insulated from the month-to-month swings of its own components — Bucharest’s sole-trader volatility and Ilfov’s incorporated-company stability offsetting each other — and a provincial top tier in which Cluj and Timiş have swapped places in a single month. Concentration at the top is stable. Ranking within it is not.