SF Is Now the Only Top-Tier Seller's Market: 9.6% Price Cuts vs 21.1% Nationally as AI Wealth Meets 7% Mortgages

Redfin's October analysis (via Fortune, Oct 6) delivers the sharpest relative-strength datapoint yet for the SF AI-housing channel: among the 50 most populous metros, San Francisco had the lowest share of active listings with a price cut — 9.6% — against a national record-for-the-season 21.1% (four weeks ending Sept 20). Fortune: "The city is home to OpenAI and Anthropic, and highly paid AI workers have been pouring their salaries, signing bonuses, and equity into local real estate" (Fortune). SF is one of just five seller's markets among major metros Redfin tracks — the others being Newark (12.2%), Chicago (13.3%), New York (13.6%), and Miami (13.7%). Note what that list is: AI/finance wealth capitals plus supply-starved Chicago and Miami.

The national backdrop makes SF's number stand out more, not less: sellers outnumbered buyers by 58% in August — the widest gap in Redfin records back to 2013 — and the Freddie Mac 30-year fixed hit 7.28% on Oct 1, the highest since November 2023, up from 6.34% a year earlier. The mechanism is now explicitly a wealth-vs-financing split: mortgage-dependent buyers are being priced out nationally while equity-rich AI workers transact anyway¹; UCSD's Michael Reher adds that rising price cuts "strengthen the case for paying cash" (his research: all-cash buyers pay ~10% less). That is the AI-wealth channel compounding at exactly the moment the national market stalls (National Housing Data Is Now a Bad Average: 4-Year-High Price Cuts and Falling Pending Sales, With the Metro Spread Running Denver 30.9% → SF 9.6%).

This partially answers the standing question of whether SF was "rolling over" (Cotality had SF −1.4% MoM / −2.6% over three months in July data): on the demand side, no — SF remains the strongest major market in the country; the monthly wobble is rate-driven, and the biggest pending wealth event, the Anthropic IPO, is still ahead (Anthropic's IPO Clock: Investor Day Oct 14, Public S-1 Due ~Oct 23, Trading Target Before Thanksgiving).

An instance of AI wealth inflates housing only where supply cannot respond. — SF's AI-wealth demand keeps it the only top-tier seller's market at 9.6% price cuts even as 7.28% mortgages stall every other major metro — AI exposure pays only where the market cannot answer it. ↩︎

Part of

This finding is an example of a pattern recurring across your work:

Central business districts starve as real estate capital shifts to AI data centers and tech-gateway housing.

AI payroll wealth converts straight into +50% neighborhood rents and IPO-cohort buying power in a supply-frozen gateway city — the hyper-inflated tech-gateway residential bubble this theme names, with Oakland's 38% office vacancy as its starving twin.

AI payrolls set San Francisco's rents; City Hall sets New York's.

The opposite pole of the same split — AI payrolls against record-low construction repricing SF rents vertically in the very quarter New York's comparable stock was capped by political price control.

Aggregate real estate averages collapse under the weight of localized technology capital.

This shows how localized surges of AI wealth insulate specific housing markets from national high-interest-rate corrections.

Paper AI wealth cannot buy houses until an IPO mints it.

The rental boom is the paycheck channel; the for-sale bid still waits on the liquidity event that converts OpenAI and Anthropic equity into house-down-payments.

Revision history

October Redfin data: SF lowest price-cut share (9.6%) and one of only five seller's markets among top-50 metros; national 21.1% at 7.28% mortgages.

4d ago· by the agent

October Redfin data: SF lowest price-cut share (9.6%) and one of only five seller's markets among top-50 metros; national 21.1% at 7.28% mortgages.

5d ago· by the agent

October Redfin data: SF lowest price-cut share (9.6%) and one of only five seller's markets among top-50 metros; national 21.1% at 7.28% mortgages.

6d ago· by the agent