Rent pressure is the clearest current tension in 94121, which is both Zillow’s ZIP market identifier and the matching Census ZCTA label. Zillow’s June 2026 ZIP ZORI, a typical observed asking-rent index blended across rental types, is $3,781 per month, up 17.2% from the same month a year earlier. Applying a 30% required-income screen to that index produces $151,240 a year, above the matched ZCTA’s ACS median household income of $133,358. That arithmetic screen places the index at 34.0% of that median; it is not advice, an applicant-qualification rule, or evidence of what any household pays. It does establish a useful separation between a current asking-rent measure and household-income context.
Comparison with the Census figure should not flatten that separation. The matched Census ZCTA’s ACS 2024 five-year survey places median gross rent at $2,363, and the current Zillow index equals 160.0% of that figure. ACS describes occupied renter homes over a five-year survey window and median gross rent includes selected utilities, whereas ZORI is a contemporaneous asking-rent index. A ZCTA is a statistical area used by Census and is not identical to a USPS delivery ZIP, even when it matches the label. The divergence therefore is source, timing, coverage, and definition context—not proof that like-for-like units changed by that amount.
Bedroom figures need an equally narrow reading. Scaling the ZIP ZORI by the local HUD ladder produces modelled monthly estimates of $2,607 for a studio, $3,123 for one bedroom, $3,781 for two, $4,830 for three, and $5,006 for four bedrooms. These are modelled estimates, not measured bedroom rents. The local HUD two-bedroom fair-market-rent standard is $3,604; HUD FMR/SAFMR is an administrative, bedroom-specific standard rather than an asking-rent observation. It supplies the scale of the model only, so the ladder cannot establish an advertised rent, utilities, condition, or unit mix for a particular listing. In particular, the matched two-bedroom figure reflects index scaling, not a direct two-bedroom rental sample.
Backward-looking history strengthens the case that the latest movement is unusually fast relative to the ZIP’s own longer path. The series has 100% coverage through its stated June endpoint. Exact same-month annualized changes were 17.2% over one year, 8.4% over three years, and 6.5% over five years. Thus, the recent direction confirms the longer rising path rather than breaking from it, but its pace accelerates markedly over the shortest window; none is a forecast or an investment recommendation. At 3.4%, annualized monthly-return variability indicates that individual monthly changes were not uniformly flat. Separately, the 11.1% maximum drawdown documents a materially larger historical retreat from a prior peak. Full coverage supports confidence that these are observed measures, yet acceleration and that drawdown limit the confidence deserved by a single current-rent snapshot. The transparent national discovery ranks among history-eligible ZIPs are 24 for momentum, 2,156 for stability, and 477 for balanced, where lower rank is higher.
Area comparisons position the ZIP between named wider benchmarks, but none replaces the ZIP series. San Francisco city, at city scope, has a rent context of $4,401; San Francisco County, at county scope, also has a rent context of $4,401; and the San Francisco-Oakland-Berkeley, CA metro, at metro scope, is $3,301. The ZIP’s current index sits below the named city and county contexts but above the named metro context. These are wider-geography context values, not local rental comps, and the county or metro figures do not alter the ACS ZCTA survey, HUD standard, or ZORI definition already used above. Scope is especially important when comparing an index with housing or income measures.
Survey housing composition adds a different constraint to the rent screen. The matched ZCTA estimates 19,777 housing units and a 7.9% vacancy rate; renters occupy 56.1% of occupied homes. Among renter households, 3,944 of 10,222, or 38.6%, are reported as rent burdened at 30% or more. The stock counts include 7,427 single-family units and 944 units in large multifamily structures. These ACS stock, tenure, vacancy, and burden aggregates describe the survey area, not real-time listings. They neither identify whether a vacancy is a rentable unit at a given price nor show that an individual renter or unit experiences the area-wide burden.
Resale, however, shows a brisk but separate for-sale signal. In Redfin’s direct rolling-three-month ZIP resale observation, median sold price was $2,257,490 and rose 19.1% year over year. The observation recorded 92 homes sold, a 14-day median marketing time, 113 active listings, an inventory count of 15 homes, and 0.5 months of supply. Average sale-to-list was 121.2%, and 79.9% of homes sold above list. Those are resale transactions and listings, not rental transactions or rental comps. The annualized ZORI divided by this median price is 2.01%, a cross-source screening ratio only—not a cap rate, net return, expected return, or property yield. Positive resale price movement directionally echoes the accelerating rent history, but the income screen challenges treating either market signal as a simple affordability conclusion.
Several boundaries remain decisive. ZORI does not supply property-specific bedroom observations, ACS does not update to the asking-rent month, HUD does not measure asking rent, and Redfin does not describe rental transactions. A property-level file would need the advertised rent and date, bedroom count, included utilities, lease term, concessions, current availability, and whether any cited sale is a genuinely comparable resale rather than merely a ZIP observation. It should also confirm the address’s geographic assignment and separate monthly rent from utilities. The published aggregates can frame those checks, but cannot resolve them or establish what a particular household qualifies for. Which verified unit-level terms remain after those source boundaries are applied?