Part 3: AI Concentrates Growth. Faropoint’s Metros Are Where It’s Going

The standard AI narrative for commercial real estate focuses on disruption: automation reduces headcounts, headcounts drive leasing, therefore AI reduces leasing demand. The math is simpler than the real economy. A more instructive question for industrial investors is where AI concentrates economic activity rather than where it eliminates it.

Oxford Economics addressed this directly in their April 2026 briefing: AI will deepen regional inequality. AI-era productivity gains will not be evenly distributed. They will compound in metros that already have dense professional services sectors, high-skill labor markets, and established financial and technology infrastructure. The chart above shows the distribution. The orange bars are Faropoint's markets.

New York, Los Angeles, South Florida, Boston, Dallas-Fort Worth, Houston, Philadelphia, Chicago, and Atlanta cluster at the top of the projected real GDP growth distribution for 2025-2030. These are not coincidentally the markets where Faropoint has built its portfolio. They are the markets that attracted the platform because of their supply-constrained industrial submarkets, dense tenant bases, and high barriers to new construction. Those same characteristics - density, infrastructure, labor depth - are exactly what Oxford's model identifies as AI-era growth concentrators.

The geographic pattern has a direct vacancy implication. In the two-way fixed-effects panel from Post 1, we found a GDP-to-vacancy coefficient of β = -0.106: each percentage point of metro real GDP growth is associated with approximately 10 basis points of small-bay vacancy tightening. If Oxford's forecast is right about where AI concentrates growth, the implication for Faropoint's markets is arithmetic, not speculative. Stronger local economies pull more service-sector demand into the infill industrial buildings that serve them.

The geographic story becomes a complete thesis when you add the tenant dimension. Oxford's sector-level GenAI productivity benefit index ranks the potential output gains from AI adoption across 20 industry categories. Transport, construction, accommodation and food services, and light manufacturing - the four sectors that dominate small-bay occupancy - all sit in the bottom third of that index. These industries rely on physical presence, tactile skills, and local relationships in ways that current AI tools cannot displace. A plumbing supply distributor is not going to automate its way out of its warehouse. A metal fabrication shop is not going to move its production floor into a language model. These are the businesses AI disrupts least, operating in the buildings AI cannot physically automate (as discussed in Post 2), in the metros where AI concentrates the economic growth that makes them more productive. Three independent lines of argument lead to the same conclusion.

Faropoint's research describes this as "double insulation": tenant insulation from AI-driven demand disruption on one axis, and geographic concentration of AI-era GDP growth on the other. Under the base case the two are reinforcing rather than offsetting — the tenant base is positioned for resilience on the downside, and the metro composition provides upside. None of this is unconditional. The whitepaper documents the horizon risks that could change this picture (autonomous trucking, AI-accelerated SMB consolidation, ecommerce channel shifts) and the framework we use to monitor them. An investor underwriting Faropoint's portfolio in 2026 is not making a bet that AI won't happen — they are making a bet that when AI happens, the buildings and markets in this portfolio are among the better-positioned places to be in industrial real estate.

Three posts so far, three independent arguments. The empirical evidence shows no AI shock in the vacancy data. The building physics keep small-bay outside the AS/RS automation story. The geography puts the portfolio in the metros where AI-era growth is heading. Post 4 turns to what we are watching on the horizon — beginning with autonomous trucking.

The complete analysis - including the Oxford Economics sector productivity data, Faropoint's full market-level regression results, and the four-point underwriting framework we derive from them - is in Supply, not AI: Why the small-bay thesis still holds — and what could change it.