Industrial Real Estate Development: Site Criteria, Tenant Requirements, and AI's Role

Industrial is coming off a historic demand cycle. The developers who succeed in this environment are the ones with the sharpest site selection and the most disciplined underwriting.

The pandemic-era demand surge pushed U.S. industrial vacancy below 3% in 2022. By early 2026, it's running at 6-7% nationally (CBRE, JLL). Still healthy, but the speculative boom is over. New starts fell roughly 40% from their 2022 peak. The market has normalized.

For developers building industrial today, the fundamentals still support well-located, well-specified projects. The question is whether your site selection discipline is sharp enough to separate the good sites from the marginal ones.

Types of Industrial Assets and Their Requirements

Industrial is not a single asset type. Each format has distinct site criteria.

Bulk distribution and logistics (500,000+ sq ft) requires 36-40 ft clear height, deep truck courts, large trailer storage, and either rail access or highway interchange proximity. These buildings are operationally demanding and tenant pools are concentrated -- Amazon, Walmart, major 3PLs, and large retailers. Spec development at this scale requires high conviction in the specific submarket.

Light manufacturing (50,000-200,000 sq ft) prioritizes power capacity, floor load, column spacing, and often proximity to a skilled labor pool. Clear height requirements are lower, but power is non-negotiable. Light manufacturing tenants are harder to screen in advance -- the asset needs to be generic enough for a range of uses.

Cold storage is the most supply-constrained industrial segment. Purpose-built cold storage requires refrigeration infrastructure, high power density, dock seal specifications to food safety standards, and site access for food-grade fleet operations. Vacancy in institutional cold storage markets is effectively zero in most cities. Development is constrained by construction cost complexity and operator concentration.

Last-mile fulfillment (under 150,000 sq ft) trades size for location. Urban and dense suburban infill sites, acceptable at lower clear heights, within short drive times of high-population-density zip codes. Land cost is the primary constraint. Amazon, UPS, FedEx, and third-party logistics providers drive most of the demand.

Core Site Criteria

Clear Height

Modern logistics tenants require 36-40 ft clear height for new construction. Legacy buildings at 24-28 ft are functionally obsolete for large-format distribution. If you're developing bulk distribution, build 40 ft. Anything less limits your exit tenant pool and your resale. You cannot retrofit clear height.

Truck Court Depth

Standard tractor-trailer operations require 130 ft minimum. High-throughput buildings spec to 185-200 ft. More truck court is always better. You also cannot retrofit this.

Power Availability

Light manufacturing and cold storage sites require 3,000-5,000 amps at 277/480 volt, three-phase. Getting that capacity from the utility can take 12-24 months. This is now one of the leading causes of industrial development delays, second only to entitlement.

Confirm available utility capacity -- not just that a line runs nearby -- before signing a purchase and sale agreement. Utilities can and do say no.

Labor Market Proximity

Distribution and fulfillment centers are labor-intensive. Tenant site selection teams run their own labor market studies: population within 30, 45, and 60 minutes of the site, competing employers, wage rates, historical turnover. If the labor market analysis doesn't work for a target tenant, the building doesn't work.

This is a site criterion that AI can now analyze programmatically, pulling labor market data, commute radius population, and employment density -- but the interpretation still requires development judgment about local conditions.

Zoning and Entitlement

Industrial zoning has tightened in many markets. Community opposition to truck traffic, noise, and light pollution is real and growing. Several large industrial projects in Southern California were killed in 2023-2024 by local opposition that wasn't adequately scoped during due diligence.

Pre-zoned industrial parks -- where the entitlement risk is already resolved -- carry a land premium but save 12-24 months of entitlement timeline. Whether that premium is worth paying depends on the land basis relative to the development returns.

Build-to-Suit vs. Spec Development

Build-to-suit means a named tenant before you break ground. Their program requirements govern the building: column grid, dock door ratio, ESFR sprinkler rating, trailer storage count, office percentage, even paint colors.

The risk in build-to-suit is concentration. You're building for one tenant on their timeline with your capital. Lease credit, lease term, and rent level all need to justify single-tenant risk. For institutional developers, build-to-suit is often preferred because the execution risk profile is cleaner -- you know what you're building and for whom.

Spec development requires deep conviction in the submarket and a building spec that matches what tenants in that market are actually looking for. The strongest spec projects are those where the developer can point to three or more active requirements in the market that the building would satisfy.

Spec development in a normalized market is not wrong. It's just less forgiving. The sites and specs that worked in 2021 do not automatically work in 2026.

AI in the Industrial Development Workflow

Site screening for industrial was historically a ground game: driving markets, calling brokers, manually cross-referencing zoning maps against MLS data and utility capacity maps. It worked. It was slow.

AI-native tools now let development teams run criteria-based site screens across multiple markets simultaneously. Inputs: zoning classification, minimum acreage, clear height potential, power availability, truck court orientation options, proximity to highway interchanges, labor market indices. Sites that clear all criteria get flagged for field review. Sites that don't get eliminated without a site visit.

Tenant spec matching runs in parallel. When a build-to-suit RFP comes in -- clear height, dock ratio, power, truck storage, rail required, timeline -- an AI system can compare those requirements against available or developable sites in the target market and return a shortlist in hours.

Permitting risk overlays add another layer: wetland delineations, FEMA flood zones, endangered species designations, community opposition signals from prior projects in the jurisdiction. This research used to take a week of consultant time. It now takes an afternoon.

The reduction in time between identifying a site and making a go/no-go decision is where AI creates the most tangible value in industrial development. In a market where the best sites get optioned quickly, speed-to-conviction is a competitive edge.

Where the Market Is Heading

Supply is moderating and demand from e-commerce and logistics is growing more slowly than during the pandemic surge, but remains positive. Structural demand from food supply chain, manufacturing reshoring, and data center-adjacent industrial provides a floor.

Cold storage and last-mile fulfillment are the highest-demand segments with the least speculative supply. Bulk distribution markets are bifurcating: submarkets with strong population growth and limited land are tight; secondary markets that overbuilt in 2021-2022 are working through excess vacancy.

For developers, the next industrial cycle will reward site quality and tenant relationships over land banking. The teams with the sharpest site selection process -- not just the most aggressive pipeline -- will build the most profitable projects.