Baku Gained More Nights Than the Regions Lost

Azerbaijan's State Tourism Agency does not publish city-level accommodation data on a dashboard. It released it in response to a press enquiry in August 2026, and the numbers describe a sharp divergence that nothing on the Azerbaijani property internet has yet absorbed.

In the first quarter of 2026, against the same quarter of 2025: Baku's overnight stays rose 31.9%. Every other major destination in the country fell, by between 32% and 60%.

The Table

Q1 2026 against Q1 2025. Revenue in thousand manat, overnight stays in nights.

City	Revenue 2026	Revenue 2025	Change	Stays 2026	Stays 2025	Change
Baku	743,021.1	543,890.7	+36.6%	11,457,571	8,684,129	+31.9%
Guba	33,541.3	47,620.5	-29.6%	465,520	712,053	-34.6%
Nakhchivan	30,296.5	47,493.3	-36.2%	388,921	616,997	-37.0%
Gusar	28,419.4	57,687.6	-50.7%	254,630	630,310	-59.6%
Gabala	27,209.3	61,596.8	-55.8%	329,902	672,983	-51.0%
Sumgayit	21,006.3	28,052.9	-25.1%	447,590	684,815	-34.6%
Absheron	20,042.3	32,889.3	-39.1%	453,642	780,163	-41.9%
Khachmaz	12,961.2	21,962.2	-41.0%	178,878	290,031	-38.3%
Sheki	12,668.1	19,250.7	-34.2%	251,931	369,412	-31.8%
Ganja	9,986.5	22,942.2	-56.5%	187,369	415,828	-54.9%

The nine regional destinations together went from 5,172,592 nights to 2,958,383 - down 42.8%.

The Finding Nobody Can Get From the Headline

The obvious reading is that visitors moved from the regions to the capital. The arithmetic does not support that, or at least not only that.

Baku's absolute gain was 2,773,442 nights. The nine regions' combined absolute loss was 2,214,209 nights. The gain is larger than the loss by about 559,000 nights. A pure reallocation would have them roughly equal.

So this is not one pie being resliced. Baku added more than the rest of the measured country shed, which means either new demand arrived in the capital, or accommodation that was previously unregistered came into the figures, or both. The agency gives no explanation and we are not going to invent one.

Why You Must Not Divide Stays by Tourists

The same release gives Baku 466,329 tourists in Q1 2026, up 9%. It is tempting to set that against 11,457,571 overnight stays. Do not.

That division yields 24.6 nights per tourist, which is not a credible average for leisure or business travel to Baku. The two series are measuring different populations - the stay count almost certainly includes domestic guests and long-stay accommodation that the tourist count excludes.

This matters because the two growth rates differ sharply: tourists +9%, overnight stays +32%. Whatever drove the stay increase, it was not mainly more arrivals. Anyone modelling short-let demand off the headline tourist figure is using the smaller and slower of the two numbers.

The Scale, Which Is Easy to Underestimate

For the full year 2025, Baku recorded 47,249,126 overnight stays. The next largest was Absheron at 3,358,485.

Baku is 14.1 times the next city, and 67% of all nights recorded across the ten largest destinations. Azerbaijan does not have a tourism market with a capital in it; it has a capital with some regional tourism attached. Every assumption about regional short-let yield has to survive that ratio first.

City	Overnight stays, full year 2025
Baku	47,249,126
Absheron	3,358,485
Nakhchivan	3,129,789
Guba	3,002,148
Sumgayit	2,802,039
Gabala	2,760,724
Gusar	2,466,353
Ismayilli	2,099,804
Khachmaz	2,034,230
Lankaran	1,565,437

What This Does to a Regional Short-Let Case

Ganja is the clearest example because it fell hardest: revenue down 56.5%, nights down 54.9% in a single year. A short-let model for Ganja built on 2024 or early-2025 occupancy is describing a market that halved. Our Ganja buying guide should be read with that in front of it.

Gabala and Gusar, the two mountain-resort destinations usually sold as the diversification play away from Baku, fell 51.0% and 59.6% in nights. Those are the places where "second home that pays for itself through holiday lets" is most often pitched.

Nakhchivan fell 37.0%, which is relevant to our Nakhchivan overview and to the health-tourism case around Duzdag - a facility whose own constraint is seasonal demand rather than capacity.

None of this says the regions are a bad place to own property. It says the rental case for them cannot be asserted from a national growth story, because for these nine places there was no growth in this quarter.

Four Caveats, and They Are Real

It is one quarter, and it is winter. Q1 covers January to March. A mountain resort's Q1 depends on snow, and a single bad season can produce a 50% swing that says nothing about the year. The full-year 2025 table above is the steadier reading.

The comparison base may be unusual. Q1 2025 could have been exceptionally strong for the regions for reasons the release does not discuss. A fall from a peak is not the same as a decline.

The source is a press enquiry. These are official State Tourism Agency figures, but they reached the public through a journalist's request rather than a published statistical release, which makes them harder to re-check and means no methodology note accompanies them.

"Overnight stays" is undefined here. As the tourist-count problem above shows, we do not know exactly what the series counts. It is internally consistent across cities and years, which is what makes the comparisons usable, but the absolute numbers should not be read as hotel-guest nights.

What to Do With It

If you are buying in Baku, this is the strongest measured support for the demand side that exists in public: a third more nights year on year, and 67% of the country's recorded accommodation demand in one city. Pair it with our price per square metre by district and the rental yield data before deciding what that demand is worth paying for.

If you are buying in a region, ask the seller for occupancy evidence from 2026 specifically - not 2024, and not a national average. Our guide to which areas suit which tourist type and the seasonal demand calendar set out what the pattern looks like when it is working.

Sources

• State Tourism Agency of the Republic of Azerbaijan, city-level tourism revenue and overnight stay figures for full-year 2025 and for Q1 2026 against Q1 2025, released in response to an enquiry from Modern.az and published by it on 21 August 2026. Read 24 October 2026. All figures in this article are from that release; the percentage changes and the combined regional totals are our arithmetic on it.
• The 466,329 Q1 2026 tourist figure for Baku and its 9% increase are from the same release.

Not verified: what the agency counts as an overnight stay, why the nine regional destinations fell, whether Q1 2025 was an unusual base, and whether any figure has been revised since 21 August 2026. No methodology note accompanied the release.