In January, the two firms most hotel investors listen to said US room rates would rise about 1 percent this year. Occupancy would slip slightly to 62.1 percent. Revenue per available room, the number the industry actually runs on, would grow 0.6 percent. It was a forecast for a flat, boring year.

Eight months in, that forecast has been revised twice and still undershoots what actually happened. We think the gap between the January number and the September reality is the most useful thing a traveler can know heading into autumn booking season.

What the numbers did

CoStar and Tourism Economics upgraded their 2026 US outlook in June, lifting the average daily rate assumption by a full percentage point, raising occupancy to 63.1 percent, and flipping a projected occupancy decline into modest growth. RevPAR growth went to roughly 2.8 percent. Full-year ADR guidance now sits near 3.1 percent, roughly three times the original call.

The actual quarterly data ran hotter than even the revision. CBRE reported that US hotel average daily rate rose 4.4 percent year over year in the second quarter of 2026, with occupancy up 0.8 percent and RevPAR up 5.7 percent.

The Labor Day weekend just past, 3 to 7 September, made the trend visible to anyone booking. Domestic hotel stays over that weekend cost about 9 percent more than the same weekend in 2025. International stays were up 12 percent.

Why the forecast missed

Forecasts of this kind are built on supply and demand. Supply was the easy half. New US hotel construction has been slow since interest rates rose, so the number of rooms coming online in 2026 was known well in advance and it was small.

Demand was the half nobody called correctly. The January models assumed a soft consumer, softer inbound international travel, and reduced government and corporate travel. Two of those three came true. What the models underweighted was that leisure demand held while room supply stayed flat, and in that arithmetic, rate does all the work.

There is a second factor that gets less attention. When occupancy is flat, hotels stop discounting. A property running at 63 percent has no reason to drop its rate to fill the last rooms, because those rooms will fill anyway. The discounting behavior that used to cap rate growth in soft years simply did not appear.

The growth is not evenly spread

This is the part that matters for where you book. The rate growth in 2026 has been concentrated in higher-tier hotels. Luxury and upper-upscale properties have carried most of the increase.

Mid-scale and upscale destination resorts and city hotels are showing occupancy or rate erosion, and the more mass-market end of luxury has been visibly soft this year. Ultra-luxury has been almost entirely insulated from the economic uncertainty everyone was worried about in January.

So the average masks a split. The $600 room went up. The $180 room mostly did not, and in some resort markets it went down.

The inbound question

One assumption in the January models did hold. International arrivals into the United States have been weaker than in 2025, and in gateway markets that matters a great deal, because inbound visitors book longer stays at higher rate tiers than domestic leisure travelers do.

What the models got wrong was the compensating effect. Domestic leisure demand absorbed the gap in most markets, and because domestic travelers book shorter and later, hotels ended up with similar occupancy on a shorter booking curve. A shorter curve gives revenue managers more room to price aggressively close to arrival, because they can see demand building rather than guessing at it months out.

The practical consequence is that the old advice about booking far ahead has weakened. In a year where hotels are pricing off live demand signals with occupancy near 63 percent, the last-minute rate is no longer reliably the worst one.

What travelers did about it

They shopped differently. One of the more striking data points out of the 2026 hotel price index work was an 1,800 percent surge in use of the budget filter on booking searches. That is not people deciding to stay somewhere cheap. That is people screening for affordability before they look at anything else, which is a behavioral shift rather than a preference shift.

The same pattern shows up globally in the price data. Vietnam was the year's standout on the upside, with hotel prices up 36 percent year over year. The UAE went the other direction, posting the steepest advertised price decline of any country for the second half of 2026 at 48.1 percent, which is what happens when a country adds a very large amount of new room supply at once.

What this means for booking this autumn

Four things follow from the data, and none of them require any faith in a forecast.

Rate, not availability, is your problem. Occupancy is only 63 percent. Rooms exist. You are competing on price, not on scarcity, and that changes the strategy. Booking earlier does not help much when supply is loose. Comparing more thoroughly does.

The mid-tier is where the value moved. If the increase is concentrated at the top, the four-star property that used to feel like a stretch has probably closed less of the gap to the five-star than the headline suggests, and the three-star has barely moved. Trading down one tier is worth more in 2026 than it was in 2024.

Watch second-half strength. Top-line performance is expected to strengthen through the back half of the year, concentrated again among higher-tier hotels. Autumn and holiday rates at upscale properties are the ones most likely to keep climbing.

Total price is doing more work than nightly rate. When base rates rise, resort fees, parking, and destination charges tend to rise with them, and they are less visible. Compare the number at checkout, not the number on the search card.

The part nobody in the industry says out loud

A 3.1 percent ADR increase in a year forecast at 1 percent is not primarily a story about demand. It is a story about pricing power. When supply is constrained and occupancy holds, hotels raise rates because they can, and the distribution layer between hotels and travelers has no incentive to push back, because most of that layer earns a percentage of the rate.

That is the structural thing we keep coming back to at Best. A booking platform paid on commission makes more money when your room costs more. We built the pricing model the other way around, so the lowest available rate is the one you see. In a year where the industry beat its own rate forecast by three times, that gap is worth more than usual. You can compare rates at best.so.

Questions we are getting about 2026 hotel rates

How much have US hotel rates gone up in 2026?
Average daily rate is tracking around 3.1 percent for the full year, against an original January forecast of 1 percent. Second-quarter ADR was up 4.4 percent year over year, with RevPAR up 5.7 percent.

Will hotel prices drop later in 2026?
Unlikely at the upper end. Industry expectations are for stronger second-half performance concentrated in higher-tier hotels. Mid-scale and resort properties are the segment showing softness, so that is where any relief appears.

Why did the 2026 hotel rate forecast get revised upward?
New room supply stayed unusually low while leisure demand held up better than the January models assumed. With occupancy steady near 63 percent, hotels stopped discounting to fill rooms, and rate carried the growth.

Where did hotel prices fall in 2026?
The UAE recorded the steepest advertised price drop of any country in the second half of the year at 48.1 percent, largely a supply story. Vietnam ran the opposite direction, up 36 percent year over year.

Related reading. What RevPAR means for the price you pay and why shoulder season is getting shorter.

Images. Hero and hotel facade via Pexels. Publix Hotel lobby, Seattle, by Joe Mabel via Wikimedia Commons, CC BY-SA 3.0.