Nikil's September 2026 South Bay STR forecast: 7 predictions, plus how August went

I'll come back on October 1 and grade myself on everything below.

Before that, August. It didn't go well, and the interesting part isn't that I got calls wrong. It's that most of them turned out to be unscoreable, which is a worse problem and entirely my own doing.

How August went

Seven calls. Two clear misses, one partial, and four I can't score.

Premium pricing for Labor Day weekend. Miss. I predicted that units priced at a 25%-plus premium for the September 5-7 weekend would be materially less booked by August 31 than units priced near a normal weekend rate. They weren't. The premium-priced units held up fine, which means I was wrong about how much resistance there'd be, and I'd rather know that than not.

PG&E shutoff before August 31. Miss. I predicted at least one PSPS warning or event touching Santa Clara or San Mateo service areas. PG&E's own PSPS page lists a single 2026 event, on July 15, and nothing in August. Fire season ran late and mild on that measure.

Leisure fade versus weeknight occupancy. Partial. The call had two halves: leisure bookings down at least 25% in the last ten days of August against the first ten, and weeknight occupancy flat or rising. The weeknight half was right, occupancy held. I never measured the leisure half properly, so I'm scoring this half-credit rather than claiming a win.

SJSU versus Stanford in the August 14-18 window. Incomplete. 30-plus day bookings at 35% of new August reservations. Incomplete. Fall pricing reset gap of 15 points. Incomplete. Pricing tools dropping September base rates 8% by August 25. Incomplete.

Four out of seven unscoreable. That's not a forecasting record, it's a bookkeeping failure.

What I'm changing

The August calls were written to sound precise. Several of them required me to run an analysis I don't run in the normal course of business, which means at the end of the month there was no number sitting anywhere to check them against.

So from here, every call has to clear one of two bars. Either it's verifiable from a public source that anyone reading this can check themselves, or it's a single number I can pull off my own dashboard in under a minute. If a prediction needs a special study to score, it doesn't go in.

That means the calls below are less dramatic than August's. They're also going to get graded, which is the point of doing this.

The seven calls for September

1. Labor Day weekend is a trough, not a peak.

South Bay demand is business-led, and holiday weekends empty the corridor rather than filling it. This is the opposite of what a coastal or mountain market does, and it catches new hosts every year.

The call: my portfolio's occupancy for the September 4-7 nights comes in below its occupancy for the following weekend, September 11-13.

2. Dreamforce helps a little, and only a little.

Dreamforce runs September 15-17 at Moscone in San Francisco. Every year some South Bay hosts convince themselves it's their week. It isn't, but it isn't nothing either, because SF rates spike and a portion of attendees price-shop southward.

The call: I take at least one booking during September 14-18 from a guest traveling to a San Francisco event, and Dreamforce-week occupancy across my portfolio does not exceed the September average by more than 10 percentage points.

3. The Stanford calendar shows up in inquiries before it shows up in bookings.

New Stanford undergrads arrive September 15, returning students September 17, and instruction begins September 22. Around that, visiting families and academics generate 30-plus day inquiries.

The call: my 30-plus day inquiry count for September 1-15 exceeds my count for August 16-31.

4. The Levi's weekend prices itself.

The Earthquakes host LAFC on September 19, and the 49ers are home against Miami on September 20 and against Arizona on September 27. Three events, two weekends, all within a short drive of a lot of Santa Clara inventory.

The call: nightly rates achieved on September 19 and 20 across units within about fifteen minutes of Levi's Stadium clear my September median nightly rate by at least 15%.

5. September steps down from August without falling off a cliff.

AirROI's San Jose data puts market-wide occupancy around 41% for the twelve months through July, with peak months near 49% and the softest near 43%. The seasonal swing here is smaller than hosts assume.

The call: my September occupancy lands below my August occupancy, and above 40%.

6. The seasonal adjustment shows up in occupancy, not in rate.

This is the one I'd most like to be right about, because it's the mistake I see hosts make every fall. They cut nightly rates hard in September and give away margin on nights that would have booked anyway.

The call: my September ADR lands within 10% of my August ADR, in either direction.

7. Zone 0 doesn't reach South Bay hosts this month.

California's Zone 0 defensible space rules were approved August 19 and are expected to take effect around the end of September. They apply in state responsibility areas and very high fire hazard severity zones, which covers very little of the flat, urban South Bay where most of my units sit.

The call: no host I work with in San Jose, Santa Clara, or Sunnyvale receives a Zone 0 notice or citation from a fire agency during September.

What could change these

A red flag event would break calls 5 and 7 at once, and would probably move 6 as well. Fire season ran quiet through August, which is not a guarantee about September and October.

A 49ers injury or a bad start could soften call 4 by late month, though the September 20 game is early enough that it should be insulated.

And if the Fed moves on September 16 in a way that rattles tech hiring sentiment, the business-travel base that holds up my weeknight occupancy is the first thing to soften. That wouldn't show up in September numbers. It would show up in October's.

How to use this

If you take one thing from the list, take call 6. Look at your September calendar and check whether you're solving a slow month by cutting rates or by improving what the platform can see about your listing. I've argued that the listing data matters more than the price now, and September is when that gets tested.

Call 1 is the practical one for this week. If you priced Labor Day weekend like a holiday, you may want to look at it again before Friday.

And if any of this has you questioning the whole model rather than the month, the STR versus lease math is the bigger question and it's worth settling before you plan another year.

Want your September and October pricing looked at by someone who does this for a living? Get a free rental analysis and I'll go through your calendar with you. Or call me at (408) 813-8001.

Sources

• San Jose, California Airbnb Data 2026 — AirROI
• Public Safety Power Shutoffs — PG&E
• Dreamforce 2026 — Salesforce
• Levi's Stadium event schedule — Ticketmaster
• Stanford Academic Calendar 2026-2027 — Stanford University
• Defensible Space Zones 0, 1 and 2 — California Board of Forestry and Fire Protection
• FOMC calendar — Federal Reserve