Robot Report Podcast: humanoids show no ROI yet despite 45 logistics pilots

On a Halloween bonus episode of The Robot Report Podcast, executive editor Steve Crow read a quote from GXO's Patrick Kelleher, given a few months earlier: GXO has run 45 humanoid pilots, has another launching in Europe, and has "not achieved ROI on humanoids yet," adding, "I think we're a couple of years away from that." GXO was the first company to deploy Agility Robotics' Digit, and Crow noted it has tested both wheeled and legged platforms.

The cost side explains the gap. Senior editor Mike Oitzman cited Agility CFO Michael Beer: a bill of materials of roughly $150,000 per robot at launch volumes of around a thousand units a year, which could fall to about $30,000 at 10,000 units a year. The hosts said Agility reported $1.8 million in 2025 revenue, that Digit V5 has been shown but is not yet available, and that expected proceeds from its SPAC listing fell from more than $620 million to $465 million. Their summary: no humanoid maker is deploying at scale yet.

Conventional automation is moving faster. Oitzman reported A3 Q3 North American robot orders up 24.8% in units year over year, just under 11,000 robots worth about $723 million, with life sciences among the segments showing gains. Cobot orders rose about 56% in units. A3 tracks industrial robots and cobots only, not mobile robots or humanoids.

For health system COOs and CFOs, the lesson is about sequencing. One of the largest logistics operators in this market has run dozens of pilots and is still waiting on payback, and the unit economics depend on volumes manufacturers have not reached. Humanoid capital requests in supply chain, pharmacy or materials handling deserve the same scrutiny GXO applies to its own, while proven categories like cobots are where orders are actually growing.

A caution on the evidence: these figures were spoken on a podcast, including secondhand quotes and projections. They are not audited results or independent studies, and the BOM path to $30,000 is a projection tied to volume.

The move: ask your operations team what return a humanoid or mobile robot pilot would need to show, and over what period, before you would fund one. Writing that threshold down first keeps a pilot from becoming a commitment, and keeps attention on automation that can relieve staff of repetitive handling work today.

You are an operations advisor to the COO and executive team of [your organization], a health and wellness organization weighing robotics in its operations.

Context: On The Robot Report Podcast, executive editor Steve Crow quoted GXO's Patrick Kelleher, who reported 45 humanoid pilots so far and no ROI on humanoids yet, which he said is a couple of years away. Senior editor Mike Oitzman cited Agility Robotics CFO Michael Beer on a bill of materials of roughly $150,000 per robot at around a thousand units a year, possibly falling to about $30,000 at 10,000 units. Oitzman also reported A3 Q3 North American robot orders up 24.8% in units, with life sciences among the segments showing gains. These are figures spoken on a podcast, including secondhand quotes and a projection tied to volume. They are not independent evidence.

Task: Prepare a question set for our operations team that establishes what return a humanoid or mobile robot pilot would need to show, and over what period, before we would fund one. Cover the specific handling work the robot would take on, the baseline we would measure against, the payback threshold, the exit criteria that keep a pilot from becoming a commitment, and how the option compares with conventional automation such as cobots.

Output: a one-page memo for the executive team with these sections: Why we are asking, Questions for the operations team, Our draft return threshold, What would stop a pilot, Open items to verify.

Use these inputs: [the areas under consideration, such as supply chain, pharmacy or materials handling], [your current annual staff hours on repetitive handling tasks], [your maximum pilot budget], [your required payback period in months].

Guardrails: Use only the facts above and the slots I fill. Cite The Robot Report Podcast by name for every figure. Flag anything uncertain or that needs verifying. Do not request or include patient data. Nothing here is medical advice. Mark the memo as a draft for the team to check.