How San Francisco Rents Have Changed Since 2015

By Homzora Team · Published September 18, 2026

A closer look at the city’s rent index, what changed over time, and how renters can use the figures without mistaking them for an apartment quote.

Data snapshot: September 18, 2026. Rent observations: January 2015 through August 2026.

San Francisco’s rental market cannot be understood through a single advertised apartment or a single dramatic headline. A listing describes one property at one moment. A historical rent series helps answer a different question: how has the broader market changed, and where does a recent observation sit within that longer history?

In Homzora’s September 2026 snapshot, the Zillow Observed Rent Index for San Francisco city reached $4,682.50 in August 2026. The corresponding figure for January 2015 was $3,062.66. Comparing those two observations produces an increase of approximately 52.9 percent in nominal dollars. That does not mean every apartment became 52.9 percent more expensive, or that every renter experienced the same increase. It describes the movement between two points in this particular index. View the published data and metadata.

The intervening years matter just as much as the endpoints. The series includes periods of rising rents, a pronounced decline around 2020, a subsequent recovery, and another period of softer values before the latest increases. Reading that sequence carefully provides a more useful foundation for a housing search than assuming that San Francisco rents move in one direction at a constant speed.

What the rent index actually measures

The Zillow Observed Rent Index, usually called ZORI, tracks asking rent changes using repeated observations of rental properties. Its methodology applies weights intended to account for differences between the properties observed on Zillow and the broader rental stock. Homzora uses the smoothed monthly city series covering all homes plus multifamily properties. Data Provided by Zillow Group. Read Zillow’s methodology.

The resulting dollar figure is an index measure of typical market rent. It is not a statement that the average current San Francisco tenant pays exactly that amount. It is also not a quote for a studio, a room in a shared apartment, or a particular building. Those distinctions are essential because readers often bring a specific housing situation to a number that summarizes a much broader market.

Consider a renter comparing two apartments with different bedroom counts, lease terms, and included utilities. The city index does not resolve those differences. It provides historical context for the search. The actual comparison still requires property details, written prices, and a clear understanding of what each monthly payment includes.

The timeline shows several different market phases

The selected observations below make the shape of the series easier to see. They are reference points rather than annual averages. August observations are useful for comparing the same calendar month across different years, while the January and December entries help illustrate other parts of the historical path.

Data Provided by Zillow Group. Figures come from Homzora’s dated city series. Values are rounded to cents.

One lesson from these observations is that a city can remain expensive while experiencing meaningful declines within a particular period. Another is that recovery does not necessarily follow a smooth path. A renter who remembers a lower quote from several years earlier may be recalling a different phase of the market, a different property type, or both.

Understanding the latest annual comparison

Between August 2025 and August 2026, the index rose from $3,739.50 to $4,682.50. The difference was $943.00, equivalent to approximately 25.2 percent. This is a comparison of the same calendar month in consecutive years, calculated from the published snapshot. It is not an annual average and should not be presented as the increase experienced by every tenant during 2026.

The arithmetic is straightforward. Subtract the earlier observation from the later observation, then divide that difference by the earlier observation. Multiplying by 100 converts the result into a percentage. Keeping the dates beside the result prevents an annual comparison from being confused with a monthly change or a forecast for the next twelve months.

A large movement deserves careful interpretation. The series establishes the observed change in the index. It does not, by itself, establish how much of that change came from employment conditions, household formation, construction, migration, or any other possible explanation. Those questions require additional evidence. Assigning a confident cause without that evidence would make the article sound more certain than the dataset allows.

Why the starting point changes the story

A comparison beginning in January 2015 answers a different question from one beginning in January 2021. The first describes a much longer period. The second begins at a lower observation within the series. Both comparisons can be mathematically correct while producing very different impressions of the market.

This is why an article should identify its comparison period before interpreting the percentage. A statement that rents have increased is incomplete unless readers know which measure is being used, where it applies, and when the comparison begins and ends. A headline based on a carefully chosen low point can exaggerate the sense that one continuous surge explains the entire history.

The figures in this article are also nominal. They have not been adjusted for inflation. The approximately 52.9 percent increase between January 2015 and August 2026 therefore describes the change in dollar terms. It does not measure the change in purchasing power, and it does not establish whether local household incomes kept pace over the same period.

A city figure cannot price a specific neighborhood

The geography in this dataset is San Francisco city. It is not the entire Bay Area, and it is not a neighborhood index. That boundary matters when someone compares San Francisco with nearby places or tries to apply the number to a search focused on one small part of the city.

Imagine two hypothetical apartments with the same advertised rent. One might include utilities and have a layout suitable for sharing. The other might require additional recurring payments and offer less usable space. A citywide index cannot determine which option better matches a household’s needs. The relevant comparison is the complete housing arrangement, including its practical limitations.

Homzora’s San Francisco neighborhood shortlist guide provides a way to organize that more specific comparison. Use the historical series to understand the market context, then narrow the search using actual listings and the features that matter to the people who will live there.

Use the trend as context for a complete budget

A housing decision becomes more informative when the monthly rent is placed beside the rest of the household budget. Utilities, transport, insurance, debt payments, childcare, and savings needs can change the practical meaning of an apparently similar rent. The index does not include a personalized calculation of those obligations.

For example, suppose a household is comparing one apartment advertised at $3,000 per month with another at $3,150. Those are illustrative figures, not market observations. If the lower advertised rent comes with substantially higher recurring expenses, the difference in total monthly cost may be smaller than the listing prices suggest. The reverse may also be true.

Upfront cash needs should be recorded separately from recurring costs. A refundable deposit, a payment made only once, and a monthly charge affect a household in different ways. Separating them makes the comparison easier to audit and prevents a temporary offer from being mistaken for a permanently lower housing cost. Homzora’s housing budget guide helps structure that exercise.

What the history cannot tell a renter

The series cannot determine whether a particular landlord will accept a lower offer. It cannot establish a legal rent increase limit for a specific tenancy. It cannot tell a household whether moving will improve its finances after all moving expenses are considered. Those are separate questions with separate evidence requirements.

It also cannot promise that the most recent pace of increase will continue. Extending a recent percentage into the future might be useful as an explicitly labeled hypothetical scenario, but it would not become a prediction simply because the starting number came from a reputable source. A historical observation and a forecast should never be presented as interchangeable.

A renter can still use the information constructively. The series may help explain why an old budget no longer produces the same search results, or why a remembered listing does not match recent options. It can support better questions. It should not replace current property comparisons or encourage a rushed decision based on the fear that a single trend must continue indefinitely.

Missing observations and revisions deserve attention

Homzora’s published series preserves missing values rather than replacing them with zeros. In this snapshot, August and September 2020 have no reported index value. A blank entry does not mean rent disappeared during those months. It means the dataset does not supply an observation that can responsibly be used in the same way as the surrounding values.

When displaying a chart, that distinction should remain visible. Drawing an uninterrupted line through missing months can suggest that every point along the path was observed. Similarly, calculating a percentage from an unavailable starting value would create an unsupported result. Good presentation makes the limits of the underlying data easier to see.

Historical values may also change when the publisher revises its series. This article refers to Homzora’s September 18, 2026 snapshot, not an automatically refreshing feed. If a later release produces a different historical value, the appropriate response is to identify the newer version and explain the revision rather than silently mixing two versions in one calculation.

Read the trend alongside other housing evidence

Rent trends become more useful when readers connect them to other questions without assuming that every dataset measures the same thing. Household cost burden describes the relationship between housing costs and income. Completed housing counts describe additions to the housing stock. Neither can be substituted directly for a market rent index.

Homzora’s San Francisco housing data library keeps those measures separate and identifies the coverage period for each. The data methodology page explains their definitions and limitations. Reading those notes before comparing numbers helps prevent a recent rent observation from being placed beside an older survey estimate as though both described the same moment.

For a household preparing to move, the practical next step is to build a current shortlist with written prices and comparable details. Keep the historical index beside that shortlist as context. The combination gives the rent trend a useful role while leaving the final decision grounded in the homes, costs, and circumstances the household actually faces.