What we’re watching next in humanoids

FedEx bets on external partners to build its warehouse-robot backbone.

In a move that reads more like a strategic pivot than a product launch, FedEx announced on March 31, 2026 that it will lean on Berkshire Gray and other external players to develop and deploy its automation tech. The company is moving from a proprietary, in-house bets-on-upgrades stance to an ecosystem play: collaborate with specialized robotics and software firms to accelerate scale across its vast network. The implication is clear: in logistics, speed and reliability now ride shotgun with access to best-in-class partners rather than a single corporate tech stack.

The shift matters for the broader humanoid and service-robot ecosystem because it addresses a core bottleneck in field deployment: complexity. Warehouse automation is not “one system fits all.” It requires a kaleidoscope of grippers, sensors, perception stacks, and control logic tuned to real-world variance—from oddly shaped items to variable box sizes and dynamic human-robot interactions on the floor. By prioritizing partnerships, FedEx signals that the hard part isn’t a shiny demo in a controlled testbed but achieving reliable, scalable performance across hundreds of facilities with predictable maintenance and supply chains for spare parts. The Berkshire Gray angle matters here: the vendor has built a portfolio of material-handling automation that aims to be modular enough to plug into existing workflows, reducing the risk of bespoke, single-vendor stacks that take ages to retrofit.

That said, this is not a magic wand. The core limitation of any partnership-led automation push is dependence on external roadmaps and service commitments. FedEx will still need to integrate vendor software with its own warehouse management and IT infrastructure, and keep humans in the loop for exception handling. The risk is not only upfront investment but ongoing alignment: software updates, hardware refresh cycles, and field maintenance across a sprawling network can erode the promised ROI if SLAs slip or spare-parts availability lags. In practice, we should expect a phased rollout with pilot sites, followed by staged scale-up—so the “field-ready” state is a moving target, not a single milestone.

Compared with a past era of internally developed pilots that struggled to leave the lab or stay on a fixed timetable, this approach is more in line with industry evidence that rapid adoption in logistics benefits from an external partner ecosystem. The hard-waked reality remains: even the most capable automation modules are only as good as their ability to harmonize with human labor, safety protocols, and the day-to-day quirks of a live network. If the Berkshire Gray collaboration translates into demonstrable gains in throughput and uptime across multiple DCs, FedEx will have proven a business model that others will copy—namely, let specialists iterate on the building blocks while the operator handles scale, integration, and governance.

What we’re watching next in humanoids

Evidence of real-world throughput and error rates at pilot sites; how quickly mispicks and jams drop with external automation partners.

Interoperability with FedEx’s legacy WMS and any standardization pushes across the vendor ecosystem; signs of convergence or vendor lock-in risk.

ROI timelines: how quickly the partnership strategy translates into lower operating costs, reduced cycle times, and capex vs. opex tradeoffs.

Maintenance and spares cadence: uptime goals, remote diagnostics, and parts provisioning across a nationwide network.

Any expansion beyond non-humanoid automation to include collaborative or humanoid-enabled tasks in loading/unloading or restocking workflows.

Sources

FedEx chooses partnerships over proprietary tech for its automation strategy