Innovation Series Warehouse Automation in 2026: What’s Real, What’s Hype and What It Means for Industrial Real Estate

Automation does not seem poised to meaningfully shrink demand for warehouse space, at least in the relatively foreseeable future. The technology with the greatest potential to reduce a warehouse’s footprint is AS/RS: automated storage and retrieval systems. By using this kind of vertical density to store more inventory in less space—picture floor-to-ceiling racking with robotic shuttles and lifts doing the retrieval work, often with few or no employees on the floor—some providers claim reductions as high as 75%. 

But AS/RS remains out of reach for most warehouse users. It’s capital-intensive, requires specific site characteristics (high clear heights, reinforced flooring, additional power), and still represents only a small share of U.S. warehouses (estimates vary depending on how “penetration” is defined, but it’s a modest slice across all sources). AMRs, meanwhile, are growing far faster but have a much smaller footprint impact. Robot-style AMRs don’t meaningfully change space needs and don’t require Class-A space or higher clear heights, while we estimate goods-to-person systems might reduce aisle space by around 15%.

Put together, even generous adoption forecasts suggest the space impact of AS/RS growth remains modest relative to the overall market—a meaningful shift for the specific buildings involved, but a rounding error against a market of billions of square feet.

Humanoid robots capture outsized attention relative to their actual deployment: Approximately 13,000 units shipped globally in 2025, according to Omdia. That being said, many humanoids in China today are shipped for research purposes, with Unitree’s IPO filings revealing that 73.6% of its humanoid revenue for the first nine months of 2025 came from “research and education.” Locomotion—getting a humanoid to walk and balance—is advancing to support a wide range of robot applications. What isn’t solved is dexterity: Robotic hands remain expensive and mechanically complex, accounting for a disproportionate share of total system cost. Battery life is another constraint, with most humanoids currently operating for only about two hours under real workloads, according to Bain, compared to industry ambitions of full eight-hour shifts. As Intrinsic CTO Brian Gerkey has put it, the hardware has largely caught up—it’s the software needed to handle real-world variability that remains the bottleneck.

China currently leads by a wide margin in humanoid hardware production. The U.S. is generally seen as ahead on the software side—the AI models that would make truly general-purpose robots possible—though most of that work still lives in research labs rather than deployed products, and it’s difficult to know how much progress may be happening less visibly elsewhere. Either way, general-purpose humanoid capability still appears to be years away from commercial reality. Compounding that, the domestic U.S. supply chain for robotic components remains extremely limited, with most parts—humanoid or otherwise—manufactured in China today. That dependency is one more reason to be measured about how quickly U.S. humanoid capabilities can scale, and it means the “humanoid workforce” narrative is likely many years away from showing up in warehouse operations at any meaningful scale.

Warehouse automation is real, accelerating in specific categories and increasingly compatible with how tenants actually operate industrial real estate—not a disruption to it. 

The technology reshaping the industry fastest is flexible, mobile automation as opposed to the vertical, space-saving systems that would meaningfully change square footage needs. That flexibility extends to leasing, too: AMRs and other modular systems can be deployed, reconfigured and removed far more easily than earlier generations of automation, making them a natural fit for standard industrial lease terms rather than a source of friction.

The automation most likely to scale in the near term can be deployed in the same Class-A and Class-B facilities that already anchor most industrial real estate portfolios, not just purpose-built mega-facilities. Because flexible automation is leading adoption while AS/RS penetration remains limited, obsolescence risk to existing warehouse stock also appears low; today’s leading automation types don’t require wholesale building replacements to stay relevant for tenants. 

The clearest financial upside of automation today accrues mainly to occupiers, not landlords. It can lower operating costs and improve throughput, largely by easing persistent labor availability and cost pressures. Whether that efficiency eventually shows up on the landlord side—e.g., through higher rents or longer lease terms—is a claim we’re not yet prepared to make. Automation adoption to date has been concentrated among the largest, highest-throughput tenants in the newest facilities, which makes it difficult to separate automation’s effects from the characteristics of the tenants and buildings that adopted it first. 

The bigger long-term variable is business model innovation—integrated fulfillment platforms that could make automation accessible to a much broader set of occupiers than today’s capital-intensive systems allow. None of this changes the more fundamental calculus for occupiers: Automation can make operations more efficient, but it doesn’t replace the need for well-located warehouse space that keeps goods close to consumers and enables rapid speed to market. 

For now, the practical takeaway of automation for warehouse landlords is site readiness: Buildings with the power, clear height and connectivity to support automation will have an edge with a growing (if still modest) slice of the tenant base, on top of the location fundamentals that show no signs of going away.