POTM Blog Issue #23, September 15, 2026
Empty Shelves
August closed faster than any August in ten years. The best explanation is not a frenzy. It is the shelf: a third fewer homes for sale than a year ago, and ten years of supply and bidding data, stretched to a shelf this empty, predict about the bids we are seeing, or a little more. Then a second question. Are prices really up 21%?
By Paulo Serna, San Francisco Real Estate Agent, Compass | Level Up Group · CA DRE# 02150409 · Published September 15, 2026
Data source Paulo’s Pulse, governed SFAR MLS data
Last week, August 2026 in Numbers recorded the fastest August in ten years: the median home went into contract in 14 days, against 29 a year earlier, and houses sold at a median 119% of asking. This issue is the meaning behind those numbers, and it fits in one word: the shelf.
San Francisco is not bidding harder than its own history says it should. It has about a third fewer homes for sale than a year ago, and the pattern of the last ten springs, stretched to a shelf this empty, predicts about the overbids we are seeing, or slightly more. Then a second question: are prices really up 21%? Mostly, but not all of it is price.
How many homes are for sale in San Francisco?
At the end of August, SFAR's InfoSparks counted 152 single-family houses for sale in the city. A year earlier it was 234, and 234 was already the lowest August of the past ten. The ten-August median is 368. Condos and townhouses: 340, against 554 a year earlier; the lowest of the ten Augusts before this one was 472, in 2018. TICs: 32, against 61.
One month can mislead, so read the year. From January through August the house shelf averaged 186 listings, against 294 over the same months of 2025: 37% fewer. Every month of 2026 so far has had fewer houses for sale than the same month of 2025, and the fullest the shelf got all year was 237, in May. Condos averaged 434 against 665, 35% fewer. TICs 51 against 80, 36% fewer. Months of supply, the time it would take to sell everything listed at the recent pace, fell from 1.3 to 0.8 for houses and from 3.1 to 1.6 for condos. This is the drought Issue #14 measured in July, still with us.
Fewer sellers, or more buyers?
A shelf empties two ways: fewer homes arrive, or buyers take them faster. All three markets lost about a third of their shelf. What happened to buying is where they part.
Houses. New contracts from January through August rose 6%, from 1,512 to 1,603, and closed sales rose 1%. In counts that is about 91 more contracts over eight months, against a shelf that averaged 108 fewer houses, and the shelf started the year already lower (172 in January against 238). These counts cannot separate fewer sellers from faster buying. My read is that both are at work: buyers are taking a few more houses, and fewer houses are sitting unsold. The result is clear either way. The houses that list sell at a median 118% of asking in 12 days.
Condos and townhouses. Contracts rose 20%, from 1,503 to 1,810, and sales 14%. Here the buyers came back in force, the condo turn this blog has followed since Issue #13. Even so, the median condo still sells at asking, in 18 days. A thinner shelf has made condos faster. It has not made them a bidding war.
TICs. Contracts up 15%, sales up 6%, 102% of asking in 22 days. It is a small market, 32 listings in August, so read the direction rather than the decimal.
Could the missing listings be selling privately instead? Some homes are. House sales recorded as sold off the MLS (the sale did not go through the MLS and was entered afterward) were 110 of 1,577 closings this year through September 11, 7.0%, against 5.6% over the same days of 2025. That share is growing, and it is worth watching. It is also about one house sale in fourteen, nowhere near enough to explain a shelf that lost more than a third.
Does the empty shelf explain the bidding?
This is the question that matters, because it separates scarcity from frenzy. If buyers were bidding beyond what supply alone explains, that would be a warning sign. So I tested it.
For every spring from 2016 to 2025, I set the median sale-to-list of the houses that closed March through June against the average months of supply that spring. The ten years fall along one curve: the emptier the shelf, the bigger the overbid, and supply alone accounts for about three quarters of the variation across those ten springs (R-squared 0.76). In plain terms, on the fitted curve, halving months of supply goes with about 18 more points of overbid.
This spring, supply averaged 1.05 months, emptier than any spring in the record. At that level the curve predicts a median of 126.2% of asking. Houses actually sold at 123.2%, on 951 sales: 3.0 points below the line, not above it.
Every spring in numbers
| Market | Spring | Months of supply | Sale-to-list | Sales |
|---|---|---|---|---|
| Houses | 2016 | 2.15 | 110.7% | 878 |
| Houses | 2017 | 1.83 | 112.9% | 884 |
| Houses | 2018 | 1.65 | 117.6% | 950 |
| Houses | 2019 | 1.88 | 111.1% | 862 |
| Houses | 2020 | 2.43 | 105.0% | 563 |
| Houses | 2021 | 1.60 | 113.1% | 1,133 |
| Houses | 2022 | 1.43 | 118.9% | 1,045 |
| Houses | 2023 | 2.00 | 105.3% | 718 |
| Houses | 2024 | 1.93 | 110.2% | 819 |
| Houses | 2025 | 1.85 | 111.1% | 871 |
| Houses | 2026 | 1.05 | 123.2% | 951 |
| Condos and townhouses | 2016 | 2.80 | 103.2% | 941 |
| Condos and townhouses | 2017 | 2.43 | 102.1% | 1,028 |
| Condos and townhouses | 2018 | 2.13 | 104.3% | 1,126 |
| Condos and townhouses | 2019 | 2.45 | 102.3% | 1,068 |
| Condos and townhouses | 2020 | 4.08 | 100.1% | 590 |
| Condos and townhouses | 2021 | 3.28 | 100.0% | 1,612 |
| Condos and townhouses | 2022 | 3.15 | 101.9% | 1,303 |
| Condos and townhouses | 2023 | 3.75 | 100.0% | 780 |
| Condos and townhouses | 2024 | 4.58 | 100.0% | 795 |
| Condos and townhouses | 2025 | 4.18 | 100.0% | 831 |
| Condos and townhouses | 2026 | 2.35 | 101.8% | 961 |
Condos tell the same story on a gentler slope. On the condo curve, halving months of supply goes with about 4 more points, not 18, which fits what condos did this year: a shelf down a third produced a faster market, not a bidding war. This spring's 2.35 months predicts 103.0%; condos sold at 101.8%, 1.2 points under.
My read, and I label it as a read: the 2026 overbid is what a shelf this empty produces, or slightly less. There is no sign of a premium on top of scarcity. The limits are real. Ten springs is a short record. One measure, supply, is doing the explaining. And 2026 sits below every supply level the ten years covered, so its prediction is an extrapolation (the dashed part of the curve). The test does not prove scarcity is the cause. What it does rule out is a 2026 overbid far above what a decade of empty shelves would predict.
Why it matters to you: in this record, bidding has tracked supply. When the shelf has been fuller, the overbid has tended to be smaller. That is a description of ten springs, not a forecast of the next one.
Is the fall listing wave arriving?
It is. POTM Command reads the MLS every day. On the August 31 drop, 140 houses were active. The count slid through the holiday weekend to 116 on September 6, then rose to 201 by September 14: 85 more houses in eight days.
How full could September get? Over the last ten years, the September shelf finished 10% to 63% above August, 34% on average. From August's 152, that arithmetic puts a normal September finish between 168 and 248 houses. The lowest September finish of those ten years was 302, in 2025. Two weeks in, the daily count sits inside the normal range, and even the top of that range would be below the emptiest September of the decade. The month is not over, and InfoSparks publishes September's count in October.
Condos show the same fall rise: 310 on September 5, 387 by September 14, still under a normal-September range of 397 to 535 and far under the decade-low September of 652.
Are San Francisco home prices really up 21%?
From January 1 through September 11, the median single-family house sold for $2,052,000, against $1,692,500 over the same days of 2025: +21.2%, on 1,577 and 1,542 sales. That is the number people quote. It is true, and it overstates what a given house gained.
A median moves for two reasons: the same kind of house costs more, or a different kind of house is selling. At our September 2 Compass citywide meeting, a colleague put median price change beside price-per-square-foot change, district by district, and made this point: where the two diverge, the mix of what sold changed. It is the right lens, so here it is on the governed record through September 11.
Price per square foot rose 15.2%, from $1,014 to $1,168. The median house that sold grew from 1,790 to 1,850 square feet, and the largest houses traded more often: 273 sales of 2,800 square feet or more, against 209. Hold the size fixed and every band gained less than the headline.
Weighted to last year's mix of sizes, same-size houses rose about 14%. Roughly two thirds of the headline is price and one third is mix. A correction to my own reading: August in Numbers called August's gain the same house costing more, because the median house sold was about the same size both years. By this measure that was too quick. Even in August the median (+24%) ran about 7 points ahead of price per foot (+17%). The median size is only one part of the mix. Over the year to date, the house gained about 14%, and the mix of what sold added the rest.
Condos have no such gap. The median condo rose 13.3% and price per square foot 12.8%, with the median size nearly unchanged, 1,161 to 1,176 square feet. For condos, the median change is the price change.
By district, the gap runs both ways.
Every district in numbers
| District (houses) | Sales 2025 / 2026 | Median price | Median change | Per sq ft change | Gap, points | Median sq ft | Share under 1,500 sq ft |
|---|---|---|---|---|---|---|---|
| Citywide | 1,542 / 1,577 | $1,692,500 to $2,052,000 | +21.2% | +15.2% | +6.0 | 1,790 to 1,850 | 34% to 31% |
| 1 Richmond | 127 / 142 | $2,315,000 to $2,615,000 | +13.0% | +11.6% | +1.3 | 2,231 to 2,289 | 18% to 11% |
| 2 Sunset, Parkside | 298 / 301 | $1,637,375 to $1,910,000 | +16.7% | +12.7% | +3.9 | 1,594 to 1,704 | 42% to 36% |
| 3 Lakeshore, Ingleside | 97 / 113 | $1,360,000 to $1,500,206 | +10.3% | +11.7% | −1.4 | 1,691 to 1,599 | 37% to 44% |
| 4 Twin Peaks West | 222 / 211 | $1,941,000 to $2,412,000 | +24.3% | +16.6% | +7.7 | 2,010 to 2,118 | 26% to 22% |
| 5 Noe, Castro, Haight | 215 / 209 | $2,555,000 to $3,200,000 | +25.2% | +27.9% | −2.7 | 2,049 to 2,078 | 22% to 21% |
| 6 Hayes Valley, NoPa | 24 / 25 | $2,900,000 to $4,090,000 | +41.0% | +37.9% | +3.2, thin | 2,615 to 2,831 | 4% to 12% |
| 7 Marina, Pacific Heights | 75 / 83 | $5,437,000 to $6,880,000 | +26.5% | +17.8% | +8.7 | 3,983 to 4,058 | 2% to 0% |
| 8 Nob Hill, Russian Hill, downtown | 17 / 26 | $3,000,000 to $4,275,000 | +42.5% | +24.0% | +18.5, thin | 2,100 to 2,680 | 29% to 12% |
| 9 SoMa, Mission, Potrero, Bernal | 166 / 168 | $1,650,302 to $2,000,000 | +21.2% | +19.4% | +1.8 | 1,630 to 1,765 | 43% to 35% |
| 10 Bayview, Excelsior | 301 / 299 | $1,075,000 to $1,180,000 | +9.8% | +11.8% | −2.0 | 1,445 to 1,450 | 53% to 54% |
In District 4 (Twin Peaks West) the median ran about 8 points ahead of price per foot as its median house sold grew from 2,010 to 2,118 square feet. District 2, the Sunset and Parkside, ran about 4 points ahead as its median house sold grew from 1,594 to 1,704. District 7 (the Marina, Pacific Heights) ran about 9 points ahead with almost no change in median size (3,983 to 4,058 square feet), so there the gap is not simple size mix; on 75 and 83 sales, which homes happened to sell matters more. District 5 (Noe, the Castro, the Haight) went the other way: price per foot rose 27.9% against a median 25.2%, so that gain is price, not mix. Districts 1, 3, 9 and 10 moved within 2 points. District 3, the Lakeshore and Ingleside side, is also where more small houses sold, 44% of sales under 1,500 square feet against 37%, which holds its median down.
The practical rule: when you price a house or judge an offer, use the price per square foot of same-size houses nearby, not the citywide median change. The median is a good thermometer and a poor ruler. It is the lesson of Issue #22 again: you are buying square feet.
What it means for you
If you are buying: the shelf is the lever. September is adding choice, and more homes to compare is the best thing that can happen to a buyer in this market. Set your ceiling from the closed price per square foot of same-size homes in the neighborhood, not from a headline median. Bidding here has tracked scarcity, so a neighborhood with more on the shelf is worth a closer look.
If you are selling: price to the per-foot sales of houses like yours, not to the 21% headline, which carries about 7 points of mix. List when the house is ready. Over the past 12 months the median house sold in 12 days at 118% of asking, and preparation, not the week you pick, is what you control.
If you are deciding whether to sell at all: nothing here says you must. If you need another season to get the house ready, or to decide, take it. These numbers describe the market; they do not make the decision for you.
And for everyone: these are citywide numbers. Your block has its own shelf. Send me an address and I will run the same cut for your neighborhood.
By the numbers
| Measure | 2025 | 2026 | Change |
|---|---|---|---|
| Houses for sale, end of August (InfoSparks) | 234 | 152 | −35% |
| Condos and townhouses for sale, end of August | 554 | 340 | −39% |
| TICs for sale, end of August | 61 | 32 | −48% |
| Houses for sale, January to August average | 294 | 186 | −37% |
| Condos for sale, January to August average | 665 | 434 | −35% |
| Months of supply in August, houses / condos | 1.3 / 3.1 | 0.8 / 1.6 | |
| New contracts, January to August, houses | 1,512 | 1,603 | +6% |
| New contracts, January to August, condos | 1,503 | 1,810 | +20% |
| Closed sales, January to August, houses / condos | 1,497 / 1,458 | 1,518 / 1,661 | +1% / +14% |
| Spring (March to June) months of supply, houses | 1.85 | 1.05 | |
| Spring median sale-to-list, houses (predicted for 2026: 126.2%) | 111.1% | 123.2% | |
| Spring median sale-to-list, condos (predicted for 2026: 103.0%) | 100.0% | 101.8% | |
| Median house price, January 1 to September 11 | $1,692,500 | $2,052,000 | +21.2% |
| Median house price per sq ft, same days | $1,014 | $1,168 | +15.2% |
| Median house size sold, same days | 1,790 sq ft | 1,850 sq ft | +3.4% |
| Median condo price / per sq ft, same days | $1,125,000 / $986 | $1,275,000 / $1,112 | +13.3% / +12.8% |
| House sales recorded as sold off the MLS, same days | 87 of 1,542 (5.6%) | 110 of 1,577 (7.0%) | |
| Houses active on the daily drop, September 6 low / September 14 | 116 / 201 |
Frequently asked questions
Why are San Francisco homes selling so fast in 2026?
Because far fewer are for sale. At the end of August 2026, SFAR InfoSparks counted 152 single-family houses and 340 condos and townhouses for sale in San Francisco, against 234 and 554 a year earlier. From January through August the house shelf averaged 186 listings, 37% fewer than 2025's 294, while new house contracts rose 6%. With 0.8 months of supply in August, the market is moving fast: over the 12 months to September 11 the median house sold in 12 days at 118% of asking.
How many homes are for sale in San Francisco right now?
On the September 14, 2026 MLS drop, 201 single-family houses, 387 condos and townhouses and 49 TICs were active (POTM Command daily count), up from a Labor Day weekend low of 116 houses on September 6. If September adds to August what the last ten Septembers did (10% to 63%), the house shelf would finish between 168 and 248, below the lowest September finish of the past decade, 302 in 2025. That is arithmetic on history, not a forecast.
Are San Francisco bidding wars bigger than low inventory explains?
Not on the evidence so far. From 2016 to 2025, the spring median sale-to-list for houses tracked months of supply closely (R-squared 0.76). At this spring's 1.05 months, that relationship predicts 126.2% of asking; houses sold at 123.2%, 3.0 points below. Condos show the same: predicted 103.0%, actual 101.8%. One caveat: 2026 supply is lower than any spring in the ten-year record, so the prediction is an extrapolation.
Are San Francisco home prices really up 21% this year?
The median single-family sale is up 21.2%, from $1,692,500 to $2,052,000 (January 1 to September 11, 2026 against the same days of 2025). About a third of that is mix: more large houses sold, and the median house sold grew from 1,790 to 1,850 square feet. Price per square foot rose 15.2%, and same-size houses rose about 14%. Condos show no such gap: median +13.3%, price per square foot +12.8%.
Are homes in San Francisco being sold off-market instead of listed?
Some are, and the share is growing, but it does not explain the empty shelf. House sales recorded as sold off the MLS were 110 of 1,577 closings from January 1 to September 11, 2026 (7.0%), against 87 of 1,542 (5.6%) over the same days of 2025. That is about one house sale in fourteen, while the number of houses for sale fell by more than a third.
Should I wait for more listings this fall before buying or selling?
Only if your own timing allows it. The fall wave is arriving (the house shelf rose from 116 to 201 in eight days), and more choice helps buyers. But even a normal September would leave the shelf below every September of the past decade, so waiting is not a reliable way to find a soft market. For sellers, nothing in these numbers says you must list this month; in a market this short of supply, a well-prepared home matters more than the week you pick.
AI Corridor Scoreboard
One reading per issue on the city's softest segment, the condos near the new AI offices, so you can watch the turn as it happens.
| Issue | Date | Reading | Call |
|---|---|---|---|
| #23 (this issue) | Sep 15, 2026 | Buyers came back and the corridor is still at asking. From June 1 to September 11, 217 condos sold in South of Market, South Beach, Mission Bay and Yerba Buena, against 151 over the same days of 2025, at a median exactly at asking (97.6% a year ago) in 25 days (49 a year ago). The share selling under asking fell from 71% to 39%. | Still the negotiator's end of the city, with less room than a year ago. |
| #22 | Sep 2, 2026 | The steepest second bedroom in the city. Corridor one-bedrooms sold at a median $702,500 and two-bedrooms at $1,350,000 over the last 12 months, a $647,500 step on 338 and 329 sales, with both rungs still at or just under asking (99.7% and 99.6%) in 34 and 23 median days. | The corridor still negotiates, and the one-bedroom negotiates longest. |
Show the 19 earlier readings
| Issue | Date | Reading | Call |
|---|---|---|---|
| #21 | Aug 30, 2026 | Still the calm end of the map. South Beach at asking on 167 sales with 29.9% over; SoMa, Mission Bay and Van Ness / Civic Center at asking; Yerba Buena 1.7% below. Meanwhile 56.6% of condos citywide now clear asking, up from 36.4% a year ago. | The negotiator's end of the market, with the floor rising underneath it. |
| #20 | Aug 25, 2026 | Two markets, one label. Corridor towers still at or below asking over the past six months (South Beach at asking on 168 sales, Yerba Buena -2.2%) while low-dues flats citywide cleared 10%+ over in 59.9% of sales. | Check the dues before you plan the offer. |
| #19 | Aug 18, 2026 | Where cash buys the most and shouts the least. Corridor condos ran 41.7% cash in the past 12 months, the heaviest concentration this series tracks, yet cash and financed closed at the same price, 99.2% versus 99.3% of asking. Cash's whole edge here is the clock: 20 median days on market against 39 financed. | Financed buyers give up nothing on price here; they pay in patience. |
| #17 | Aug 9, 2026 | Zero, which is the cleanest reading this scoreboard has produced. Across 368 corridor condo closings so far in 2026, not one sold $1,000,000 over asking, and the typical one closed slightly under asking: South Beach 0.7% below on 183 sales, South of Market 1.0% below on 79, Yerba Buena 2.4% below on 53, Mission Bay exactly at asking on 53. In an issue about where the overbidding is, the corridor is where it is not. | Buyer opening holds, and this issue measures exactly how wide it is. |
| #16 | Aug 6, 2026 | Flat in every window, which this issue argues is the whole point. South Beach condos read 100.0% of asking at 30, 90 and 180 days and 99.0% over the trailing year. South of Market reads 99.4, 99.8, 99.6 and 99.0. Four windows, one answer. Every house lane in this issue moved when the window moved; the corridor did not, and a segment that reads the same no matter how you slice it is a segment that is not turning. | Buyer opening holds, and now it is measured four ways instead of one. |
| #15 | Aug 2, 2026 | Unchanged at the bottom of the same ladder. This issue extended the over-asking gradient into small multifamily, and the corridor still anchors the low end: right at asking, while two-unit buildings cleared 11.7% over and houses 23.8%. The ordering is by how much a property lives like a house, and a corridor tower is the furthest thing from one. | Buyer opening holds, and this issue explains why it persists. |
| #14 | Jul 29, 2026 | Read on supply this time, not price. Citywide condo and townhome months of supply fell from 3.9 to 1.7 in a year and active listings from 772 to 480, so even the calmest lane in the city now offers less to choose from. This issue did not re-measure corridor pricing. | Buyer opening holds on price; the shelf behind it is thinner. |
| #13 | Jul 25, 2026 | Graduated to the essay. Back at asking for the first time since 2022, after three springs about 1% below it, and the clock changed: median market time fell from 38 days to 19 and the share selling over asking roughly doubled, from the low twenties to the mid forties. Price at par, speed doubled. | Negotiating room intact; the window now narrows in speed, not price. |
| #12 | Jul 21, 2026 | Still the calm corner, and it proves the point. District 9 condos, SoMa, Mission Bay, and South Beach, sold right at list, about 100%, with only 31% over asking on 836 sales, while mid-priced houses cleared 123 to 127% of list. Condos rarely get listed low to start a war, so the overbid never appears. | Buyer leverage holds where the list-low tactic is not used. |
| #11 | Jul 17, 2026 | Unmoved by the house story. While overbidding ran one tier below the trophy core, the corridor and condo core cleared near asking, and the two flat lanes the field named as spillover candidates, Hayes Valley and Lower Pacific Heights, stayed calm in closed data. | Still the clearest buyer opening in the city. |
| #10 | Jul 14, 2026 | Still soft while the headline is elsewhere. June's million-over-asking story is a west-side and central house market, not the AI-corridor towers. District 9 condos, SoMa, Mission Bay and South Beach, ran about 10% below last year even as volume climbed. Activity returns to the corridor; pricing has not. | Still the clearest buyer opening in the city. |
| #09 | Jul 10, 2026 | Still the soft floor at the halfway mark. District 9 condos, SoMa, Mission Bay, and South Beach, sold near 100% of list with only about 36% over asking on roughly 290 sales this year, while citywide houses ran near 121% of list. The widest lane in the city stays open. | Clearest buyer opportunity holds into the second half. |
| #08 | Jul 5, 2026 | Still the soft floor even as the top books records. District 9 condos, SoMa, Mission Bay, and South Beach, sold near 100% of list with heavy cash and light competition, while $5M+ houses set a decade volume record at about 112% of list on roughly 64% cash. Cash without a crowd here. | Buyer opportunity intact where the crowds are not. |
| #06 | Jun 25, 2026 | Still the calm corner while the house middle runs hot. District 9 condos, SoMa, Mission Bay, and South Beach, sold right at list, about 100%, with only 35% over asking on 300 sales, against the $1.5M to $3M house band at 122 to 125% of list. | Buyer opportunity holds where the bidding wars are not. |
| #05 | Jun 21, 2026 | Cash, not heat. Corridor condos carry heavier cash than the citywide condo average, about 42% versus 37%, yet still sell near 98.7% of list with only about 20% over asking versus 45% citywide. Cash concentrates here; competition does not. | Negotiating room for financed buyers. |
| #04 | Jun 17, 2026 | Still the soft floor while houses raced ahead. Corridor near 98 to 99% of list versus 103.8% citywide and about 123% for single-family in the last 30 days. | Buyer opportunity holds; the gap to houses only widened. |
| #03 | Jun 13, 2026 | Still the bottom of the overbid table. Corridor sale-to-list at about 98 to 99% versus 103.6% citywide, trailing year. | Opportunity intact for negotiators. |
| #02 | Jun 10, 2026 | Turning at the edges. Citywide condos hit 101.4% of list in May; inventory fell to 584 from 905. The corridor towers remain the soft end. | Window narrowing, not closed. |
| #01 | Jun 7, 2026 | Soft. Only 37 to 43% of SoMa, Mission Bay, and downtown condos sold over asking. | Clearest buyer opportunity in the city. |
- The shelf: 152 houses and 340 condos and townhouses were for sale at the end of August, against 234 and 554 a year earlier. From January through August the house shelf averaged 186 listings against 294 (−37%), condos 434 against 665 (−35%), TICs 51 against 80 (−36%). Every month of 2026 so far has had fewer houses for sale than the same month of 2025.
- Buying split the markets: house contracts rose 6% and sales 1%, on a shelf that started the year already lower; condo contracts rose 20% and sales 14%, as buyers came back in force. The counts cannot separate fewer house sellers from faster buying (my read: both). Houses sell at 118% of asking in 12 days; condos still sell at asking, in 18.
- The scarcity test: across 2016 to 2025, spring sale-to-list for houses tracked months of supply closely (R-squared 0.76). This spring's 1.05 months predicts 126.2% of asking; houses sold at 123.2%. Condos: predicted 103.0%, actual 101.8%. My read, consistent with and not proof of scarcity: the overbid is what the shelf would produce, or slightly less. On the fitted curves, halving months of supply goes with about 18 more points of overbid for houses and about 4 for condos.
- The fall wave is arriving: 116 houses on the September 6 drop, 201 by September 14. A September that adds what the last ten added (10% to 63%) would finish between 168 and 248, still below the lowest September finish of the decade, 302. Arithmetic on history, not a forecast.
- The 21% question: the median house rose 21.2% ($1,692,500 to $2,052,000) but price per square foot rose 15.2% and same-size houses about 14%, because larger houses sold (2,800+ sq ft sales 209 to 273). About a third of the headline is mix. Condos have no such gap: +13.3% median, +12.8% per foot. Price a home from same-size sales nearby, not from the citywide median.
Median price, sale-to-list and market time for every San Francisco neighborhood are live in the market explorer, refreshed weekly. Check your own neighborhood's shelf there before you set a budget or a list price.
Sources and further reading
- California Association of Realtors county sales activity, including San Francisco County
- August 2026 in Numbers, the record this issue explains
- Issue #14: The Drought, the supply story in July
- Issue #22: What a Bedroom Costs, why you are buying square feet
- Selling in San Francisco, starting with whether you should
- How I read the San Francisco market, the method behind these figures
Methodology and sources
Sources, each with its own date. (1) SFAR InfoSparks, San Francisco County, monthly series in the POTM Command workbook, data as of September 1, 2026: homes for sale at month end, months of supply as published, new pendings and closed sales summed across the ten districts. InfoSparks moves by a few listings between reads as late reports arrive; August in Numbers used a live read on September 5 (151 houses, 338 condos), and this issue uses the monthly series (152 and 340) so every year is read the same way, which is also why the year-over-year changes differ slightly (last week: houses -37%, condos -38%; here: -35% and -39%). (2) POTM Command governed MLS closings through September 11, 2026, 62,877 governed closed sales since 2016, parcel-deduplicated, sale-to-list plausibility band applied; medians only, never averages; price per square foot is the median of each sale's own ratio, and sales without a reported size are left out of size figures. (3) The POTM Command daily MLS drops, August 3 to September 14, 2026, counting distinct listings with an active status on each full drop day; on August 31 it ran about 8% under InfoSparks for houses (140 against 152). The spring test pairs the plain median sale-to-list of March through June closings with the average of the four monthly months-of-supply readings, fits 2016 to 2025, and places 2026 on the fit; 2026 supply is below the fitted range, so its prediction is an extrapolation. These figures were first assembled for a Compass office presentation and recomputed independently from the raw exports before publication. "Sold off the MLS" marks a sale that did not transact through the MLS and was entered afterward; how much exposure those homes had varies and is not visible in the data. Single-family, condominium and townhouse, and TIC figures are kept separate; 2-4 unit buildings are excluded. General information, not a forecast, a valuation of any specific home, or legal, lending or tax advice.
What does this Pulse mean for your block?
Two homes five blocks apart can carry very different risk. Let's talk about your specific segment, no pressure.