Mid-Tier Home Prices Fell in 24 of 33 Big Cities. San Francisco Spiked 11.6%.

Wolf Street's July 2026 update on mid-tier home prices in 33 big and expensive US cities shows year-over-year declines in 24 of them, led by Austin at -4.4%, Las Vegas at -3.1% and Nashville at -3.0% ^[1][2]. In the same month, San Francisco mid-tier prices rose 1.5% from June, taking the year-over-year gain to 11.6% ^[3]. That is 73% of the tracked cities going one way and the AI capital going the other, which means a national housing view is no longer a view at all ^[8].

The peak comparison is worse than the annual one. As of June, prices were below their prior-year peaks in 28 of the 33 cities, led by Austin at -27%, Oakland at -25% and New Orleans at -19% ^[4]. Those peaks are mostly stale: 17 cities topped out in 2022 and nine in 2024, with Boston peaking in April 2025 and San Jose in January 2025 ^[5]. Only Chicago and New York City reached new highs ^[6], which means five cities are not below peak but just two are making progress ^[7].

San Francisco is the outlier with a story attached. Wolf Richter writes that AI mania started in the luxury market, produced what is called a mansion shortage, and in recent months trickled down into the mid-tier segment ^[9]. The city was recently near the top of the decline list before flipping from falling to surging in the second half of last year ^[10]. Even after the run, July prices were still 6% below the 2022 all-time high ^[11]. Richter says a new high before yearend is possible if the mania lasts ^[12]; arithmetically it takes about 6.4% more, or roughly four months at July's 1.5% monthly pace ^[13].

Boston is the reminder that peaks can be recent and still broken: down 0.5% in July from June, down 1.4% year-over-year, and down 2.3% from its April 2025 high ^[14].

The base effect underneath all of this is enormous. Between mid-2020 and mid-2022, prices rose 62% in Austin, 60% in Phoenix, 50% in Fort Worth, 49% in Raleigh and 39% in Sacramento ^[15]. Richter attributes that to the Fed's purchases of Treasuries and mortgage-backed securities with newly created money, which produced below-3% mortgage rates while inflation surged toward 9%, and to buyers who trampled each other to lock in those rates ^[16][17]. Austin's 27% decline from peak still leaves prices roughly 18% above mid-2020 on those figures ^[22]. Owners in the fallen cities hold sub-3% mortgages on homes worth less than they paid, and Richter's point is that they are fine as long as they stay put ^[18], which tells you something about who is left in the seller pool.

Two methodological cautions before anyone trades this. The index is the seasonally adjusted three-month-average mid-tier Zillow Home Value Index for single-family homes, condos and co-ops, a backward-looking measure of the middle third by price that includes off-market and for-sale-by-owner deals ^[19]. Three-month averaging means San Francisco's spike is partly already history. And the list is deliberately skewed to expensive markets: a city needs a mid-tier ZHVI of at least $300,000 at some point, with Houston and Philadelphia admitted only because they are the fourth and sixth largest cities ^[20]. In the cities that do not qualify, five mid-tier homes together cost less than one in San Francisco ^[21].

Watch whether San Francisco prints a new high before yearend ^[12], whether the AI bid reaches Oakland, still 25% below peak, and San Jose, whose peak was January 2025 ^[4][5], and whether Austin's annual decline of 4.4% keeps steepening off a 27% drawdown ^[2][4].