More or Less?

Two truths are emerging in the San Francisco office market.

The first is surging demand for space from AI companies. It is increasingly clear that San Francisco is, and will remain, the epicenter of AI. The critical ecosystem has already been established. Talent, venture capital, strategic partnerships, M&A activity, research institutions, and universities are all concentrated here. And these companies are growing fast.

The nature of AI work, interdisciplinary, fast-paced, iterative, and highly confidential, tends to favor an in-office culture. Job growth in the AI sector has been robust. While difficult to measure precisely, estimates suggest AI-related employment in the San Francisco region has grown at an annual rate approaching 40% since 2022. AI demand has dominated the office market over the past two years and, at least for now, shows no signs of slowing.

The second truth is that AI is also causing layoffs.

We are already seeing an acceleration in workforce reductions tied directly to AI-driven productivity gains. Over the past year alone, the rapid development and deployment of agentic AI has begun reshaping how companies execute work, often reducing the need for human labor in the process.

How will this play out over time?

Our view is that the regional office market will increasingly divide into two distinct categories: high-performance assets and economically obsolete assets.

To be sure, this is already happening. Demand is concentrated in key clusters and focused heavily on higher-quality buildings. Companies, both AI firms and traditional occupiers, are competing for a much smaller pool of truly desirable space. As a result, premium rents are holding firm despite historically high overall vacancy.

In fact, today’s market often behaves less like a market with 30% vacancy and more like one with sub-10% vacancy, because so much of the inventory is effectively outside the competitive set.

Over time, if and as the desirable portion of the market fills, some demand will inevitably spill into secondary locations and lower-tier inventory. This has historically been part of the dynamic driving Oakland’s cyclical performance. San Francisco becomes too expensive, and some tenants move across the bay in search of value.

The question is not what happens if demand outpaces desirable supply. We know how that story ends.
The larger question is whether AI job creation will ultimately outpace AI-driven job loss.

Early indications are cause for caution.

We see potential disruption across a broad range of white-collar professions that currently fill San Francisco office buildings. Software engineering, legal services, finance, administrative support, consulting, and middle management all appear exposed to varying degrees of automation and workforce compression.

AI may prove to be the most significant economic change agent of modern times, perhaps unlike anything we have previously experienced.
For now, the answer remains uncertain.

Will AI ultimately create more jobs, or fewer?