The consensus read on San Francisco’s economy is that the San Francisco tech job losses don’t really matter because AI wealth is filling the gap. The payroll data from the Bureau of Labor Statistics suggests that framing may be doing a lot of heavy lifting.
What the San Francisco Tech Job Losses Actually Look Like
In San Francisco and San Mateo County, employment in “Professional, Scientific and Technical Services” dropped to 198,800 in August, the lowest since March 2019 and below the lockdown lows. The sector has shed 35,500 jobs since its peak in June 2022, a decline of 13.2%. During the Dotcom Bust, that same sector lost 28% of its employment through to October 2003. The percentage decline through August is nearly halfway there. In raw job numbers, it is all the way there.
The “Information” sector, where large tech companies, social media platforms and software publishers sit, has fared worse in proportional terms. Jobs fell to 102,530 in August, the lowest since October 2019, having shed 28,900 positions, or 22%, since the hiring peak in August 2022. The Dotcom Bust eventually took 46% of Information jobs in this area, bottoming out in mid-2006 after a multi-year industry depression. Again, the percentage loss through August is approaching halfway, but the absolute number of jobs destroyed is already comparable.
That distinction matters. The Dotcom Bust unfolded over years; the current contraction has achieved similar numerical damage in a shorter cycle, while being broadly described as a period of AI-driven prosperity.
Finance Quietly Falls to a Decade Low
Less covered than the tech figures, but arguably more telling, is the collapse in “Financial Activities” payrolls. That category, which covers finance, insurance, real estate and leasing, dropped to 72,500 jobs in August, the lowest since January 2014. Since mid-2022, the sector has lost 13,780 jobs, or 16%. The longer-term picture is bleaker still: payrolls are down 26% from the absolute peak recorded in 2001, a decline that predates and then absorbed the departures of several major employers. Schwab relocated its headquarters from San Francisco to Texas in 2019. Wells Fargo’s executive headquarters, where operational decisions are made, moved to New York City in the same year.
The financial sector’s trajectory is not a pandemic distortion. It is a structural erosion running over two decades, now accelerating.
The AI Wealth Effect Is Real But Narrow
None of this is to say the AI boom is fiction. OpenAI and Anthropic, both headquartered in San Francisco, have hired aggressively and delivered large salary and stock compensation packages. Anthropic is reportedly targeting a $2 trillion valuation for its IPO; OpenAI is said to be targeting a valuation in the $1.2 to $1.5 trillion range. Countless smaller AI-related companies have added to the hiring picture.
The problem is scale. AI-related jobs remain a small subset of total payrolls. The sector concentration that made San Francisco extraordinary during the tech boom, with Information alone accounting for 9.2% of local payrolls versus roughly 2% nationally, now works against it. A sector that is outsized in good times produces outsized damage when it contracts. The AI hiring wave has not come close to absorbing the losses in the broader tech and professional services categories it sits within.
Total private-sector payrolls in the two-county area edged up by just 7,000 jobs year-over-year in August, or 0.7%. Since the February 2020 peak, private-sector payrolls remain down by 87,600 jobs, or 8.2%. For context, the Dotcom Bust took 17% of private-sector payroll jobs in the area at its worst.
The sectors genuinely growing are healthcare services and private education, which rose to 170,400 jobs in August, up 3.3% year-over-year, and leisure and hospitality, up 5.4% over the same period to 132,400. Leisure and hospitality, however, remains 9.5% below its February 2020 level. These are not the sectors sustaining $3 million median home prices.
The consensus may be overweighting the wealth effect from a small number of extremely well-compensated AI workers and underweighting the structural employment base being quietly hollowed out beneath it. As Wolf Street noted in publishing this data, the increasing use of AI across businesses may itself be accelerating job losses in the very sectors that once defined this regional economy. The San Francisco tech job losses, viewed through that lens, are not a transitional dip. They may be a structural feature of the AI era, not a temporary side effect of it. Wells Fargo‘s own executive operations having quietly shifted to New York City captures the direction of travel rather neatly.
