At $5,472 per month in June 2026, ZIP 94114’s Zillow Observed Rent Index (ZORI) frames the current rent signal as a typical observed asking-rent index blended across rental types, rather than a quote for a uniform apartment or a lease executed this month. The source definition is central: ZORI tracks asking rents, not household rent payments, utilities, or a bedroom-specific inventory slice. The five-digit label is both a Zillow ZIP market identifier and a Census ZCTA match. A ZCTA is a statistical area and is not identical to a USPS delivery ZIP. That geographic match permits comparison, but it does not erase differences in sample, timing, or measure.
The matched Census ZCTA’s ACS 2024 five-year estimate places median gross rent at $3,071, a survey measure of occupied renter homes that includes selected utilities. The ZORI is 78.2% above it, a gap which should not be treated as error or a rent-growth calculation because these series cover different universes and periods. ACS reports median household income of $204,134. Applying the 30% screen to the current ZORI produces required income of $218,880 and an asking-rent-to-income result of 32.2%. That is arithmetic, not advice, an applicant qualification rule, or proof of what any household pays. The ACS five-year survey’s population and utility treatment also prevent it from serving as a direct substitute for Zillow’s current asking-rent index.
Backward-looking rather than forward-looking, the ZIP ZORI history through its stated June endpoint shows annualized exact same-month increases of 18.7% over one year, 8.9% over three years, and 7.4% over five years. Recent direction therefore confirms the longer positive path, but at a materially faster recent pace; that description is measurement, not a forecast or investment recommendation. History coverage is 99.1%, so the series is nearly complete for its available window. Its annualized monthly-return variability, at 3.2%, indicates relatively contained month-to-month changes by that metric. Yet a historical peak-to-trough drawdown of 12.4% shows that a current snapshot can still sit meaningfully away from an earlier level. Momentum, stability, and balanced national discovery ranks of 20, 1,932, and 359 among history-eligible ZIPs are transparent sorting tools, with lower ranks higher, not predictive grades.
Bedroom detail requires a different lens. The FY2026 local HUD FMR/SAFMR ladder is an administrative, bedroom-specific standard, not asking rent, and it runs from $2,485 for a studio to $4,772 for four bedrooms. Scaling the ZIP ZORI by that local HUD ladder produces modelled monthly ZIP estimates of $3,773 for a studio, $4,520 for one bedroom, $5,472 for two bedrooms, $6,990 for three bedrooms, and $7,245 for four bedrooms. These are modelled estimates, never measured bedroom rents or unit-level comparables. They preserve the area’s HUD bedroom relationships while anchoring the overall level to Zillow’s blended index, so they cannot establish the asking price, utility package, condition, or lease availability of a particular apartment.
Survey housing counts add another tension rather than a verdict on availability. The ACS ZCTA contains 18,771 housing units, with a 10.2% vacancy rate; it records 9,238 renter-occupied homes, or a 54.8% renter share. Of the renter households in the burden tabulation, 2,806—or 30.4%—had gross rent at or above the burden threshold. This burden result describes surveyed households, not an individual renter’s budget or the affordability of a listed unit. The stock table identifies 5,060 single-family units and 1,669 units in large multifamily structures, categories that help describe the ZCTA’s housing mix without converting either category into a rental-comparable set. Nor does the aggregate vacancy rate prove that a specific vacant home is rentable, habitable, or currently offered.
Broader geography supplies context only. In San Francisco city context and San Francisco County context, the reported rent measure is $4,401, below the ZIP index; in San Francisco–Oakland–Berkeley, CA metro context, the reported rent measure is $3,301 and the metro rent-to-income measure is 29.1%. These city, county, and metro figures are not ZIP observations and cannot replace either the ZIP ZORI or the matched-ZCTA ACS results. They nevertheless sharpen the comparison: the ZIP’s current asking-rent index is above all three wider rent benchmarks, while its simple income screen is above the metro context figure. Wider geographic context establishes relative position only, not a direct substitute for a ZIP-level rental comp set.
Resale evidence is notably brisk, but it is a different market. Redfin’s direct rolling-three-month ZIP resale observation through June reports a median sold price of $2,209,501, up 16.3% year over year, from 149 homes sold; median marketing time was 14 days. The inventory measure was 33 homes and months of supply stood at 0.7. Sale-to-list signals stayed in the for-sale universe: the average sale-to-list ratio was 119.96%, and 78.7% of sold homes closed above list. Those measures describe ZIP resale pricing and liquidity, not rental transactions, rental comparables, or property-level operating economics. The concurrent resale price change and rent-history acceleration point in the same observed direction, but neither establishes why the other moved or what happens next.
Annualized ZIP ZORI divided by the Redfin median sold price is 2.97%, a cross-source screening ratio only; it is not a cap rate, net return, expected return, property yield, or a substitute for expenses and financing. The sharp resale signals confirm the history’s recent upward direction, yet the rent-versus-income screen and the large ACS/Zillow level difference challenge any simple claim that the current index represents broadly affordable occupied rent. Both tensions are descriptive, not causal. Property-specific work would need verified comparable asking rents by bedroom, lease term and included utilities; current unit availability and condition; building type; and the actual listing, sale, and operating-cost records. What do those unit-level documents show relative to the modelled ladder and today’s index?