AI boom is city’s weirdest tech boom, says S.F.’s chief economist

San Francisco’s AI boom has attracted about 60% of U.S. AI venture capital over the past three years, or nearly $190 billion last year across 2,500 startups, but it is coinciding with a 7.5 million square foot net loss of occupied office space and more than 30,000 tech job losses over the same period. The city also faces a $643 million two-year deficit, and Mayor Daniel Lurie has proposed $400 million of spending cuts this year, including $100 million from position eliminations and layoffs. The article frames AI as a long-term economic positive for the city, but near-term budget and labor-market conditions remain weak.

Analysis

The market is likely underestimating the divergence between capital formation and local operating activity. A region can absorb a huge amount of VC and still become a net loser for traditional urban economic multipliers if the spend is increasingly remote, software-heavy, and distributed rather than office- and headcount-intensive. That implies a weaker relationship between “AI leadership” and the usual beneficiaries of a tech boom: Class A office landlords, downtown retail, and local payroll-tax sensitive municipal revenues may stay structurally impaired even if startup funding remains robust.

The second-order winner is not necessarily the city itself but the ecosystem around remote talent concentration: residential real estate, premium rentals, and select consumer services can tighten even as office utilization stays weak. That creates a bifurcated local economy where housing-linked assets outperform commercial real estate, and where the conventional recovery trade built on office reopening is the wrong expression. The most durable public-sector beneficiary is likely transit, not because it solves the deficit immediately, but because the fiscal model only works if people keep commuting into the core.

The key risk is timing: the AI capital cycle can mask structural weakness for months, while budget stress and layoffs hit faster and more visibly. If VC funding slows, or if AI hiring normalizes outside the city, San Francisco could face a simultaneous decline in high-income demand and a sharper fiscal squeeze, which would force bigger service cuts than currently priced. Conversely, if the city successfully converts AI capital into permanent residency and incremental local employment, the pessimism around municipal recovery may be overdone.

The contrarian view is that the “AI boom” may be less a broad growth regime than a narrow capital-market phenomenon. That makes the urban recovery trade fragile: headline investment can coexist with weak jobs, weak office demand, and recurring budget gaps. The right posture is to separate beneficiaries of venture asset inflation from beneficiaries of local economic throughput.

Key Decisions for Investors

• Short regional office exposure vs. long residential exposure: pair short SL Green/office-heavy proxies (e.g., SLG) against long multifamily/residential landlords with Bay Area exposure (e.g., ESS, AVB) over 6-12 months; thesis is capital inflow supports housing more than downtown office utilization.
• Underweight San Francisco-linked municipal recovery proxies; avoid chasing any rebound in local consumer/retail names tied to weekday office traffic for the next 3-6 months, as payroll and foot-traffic elasticity remains negative.
• Go long transit-enabler exposure on weakness where available: if ballot measures pass, use a tactical long in infrastructure/engineering beneficiaries; if not, fade any rally in downtown-exposed REITs and retail. Risk/reward favors waiting for the vote and then trading the policy outcome rather than the narrative.
• For tech beta, prefer venture-adjacent AI infrastructure names over local-services beneficiaries; own semis/cloud infrastructure on pullbacks, but avoid assuming San Francisco housing/retail captures the full upside of AI spending.
• If municipal deficit headlines intensify, consider a short in BBB-rated California local government-adjacent credit proxies or a hedge via muni CDS where available; the catalyst is a 1-3 month budget cycle, not a multi-year story.