IDTechEx May 19 Report: Humanoid Robot ASP Falls from $114,700 「2024」 to ~$37,000 「2030」, Sub-$5/Hour Operating Cost in High-Utilization Industrial Settings by ~2030, 6-Month Payback Today Under High Utilization vs 15 Months at Medium, Annual Shipments to Approach 1.8M by 2036 — The 「Cost Per Productive Hour」 Curve That Now Has the Standard Chartered Number Sitting Underneath It | LostJobs.AI

On the same day Standard Chartered named the year and the number for AI-driven bank layoffs, an analyst report dropped quietly out of Cambridge with the cost curve sitting underneath it. IDTechEx published Humanoid Robots: Market, Technologies, and Opportunities 2026-2036, and the numbers are concrete enough that the conversation around humanoid economics can finally stop being a vibes argument.

The headline figure is the price line. Average selling price of a humanoid is $114,700 in 2024, projected to fall to ~$37,000 by 2030 — a 68% reduction in six years, with further declines into the mid-2030s. The price curve is what unlocks everything else.

The number that matters: $5/hour by ~2030

Hardware ASP is the column most analysts cite, but it is not the operating column that drives a CFO’s decision. The number that drives the decision is cost per productive hour, and IDTechEx now has a model for it. Under high-utilization industrial scenarios, the report projects humanoid robot operating costs to fall below $5/hour by around 2030, with further declines toward 2036.

For comparison: the total employer cost of a US manufacturing worker — wages plus benefits, healthcare, payroll tax — is currently in the $40-65/hour range and rising. A Chinese factory worker is in the $5-10/hour range and rising faster. By 2030, a high-utilization humanoid is projected to undercut the lower bound of the Chinese cost line, not the US one. That is the asymmetric piece. The cost arbitrage closes in the high-cost markets first because the human worker is more expensive there. It then closes in the low-cost markets — which is when the global labor map gets redrawn.

Payback today: 6 months high, 15 months medium

The number with the most immediate practical effect is the payback period under 2026 conditions. IDTechEx puts it at ~6 months under high-utilization scenarios and ~15 months under medium utilization. That is not a 2030 forecast — that is now.

Six-month payback at high utilization is the threshold corporate procurement teams have been waiting for. At that level, a humanoid is roughly priced like an industrial machine tool, not a capital pet project. A CFO can write a check, depreciate it on a 36-month schedule, and book the labor cost savings in the same quarter. The 15-month medium-utilization number is closer to “experimental purchase” territory — but is still a faster payback than most CAPEX line items.

The framing the report is careful to land is that payback is highly scenario-dependent. A humanoid running 18 hours a day on box-handling in a Schaeffler factory has different economics than the same humanoid running 6 hours a day in a less structured Korean retail pilot. The model variables that matter are utilization, task continuity, environment complexity, and integration quality. Hardware cost is the easiest variable; the others are still being figured out.

The 1.8M shipments by 2036 forecast

The market sizing is the column most analysts will quote and most equity desks will be modeling against this week. IDTechEx projects:

Market size: ~$25 billion by the early 2030s, before moderating as the market matures.
Annual shipments: ~1.8 million units per year by 2036.
Demand structure: automotive manufacturing as the core demand base, logistics following, home-use a longer-term opportunity with limited penetration within the forecast period.

The shape of the curve is the part worth dwelling on. Automotive is leading because automotive has standardized workflows, clear task boundaries, and strong labour-cost pressure — the conditions that make humanoid deployment economically defensible. Logistics is second because warehouse handling has the same structural traits. Home use is the long tail because consumer environments are unstructured, safety-critical, and labour-cost-sensitive in a different way.

This is the same pattern that has played out for every major capital-equipment cycle in manufacturing history. The factory floor goes first; the home goes last.

The bottleneck is not hardware

The line that elevates the report from “another bullish humanoid pitch deck” is the bottleneck call. IDTechEx is explicit: the constraint on large-scale adoption is no longer hardware cost, it is effective output. That is, whether the robot can perform economically valuable tasks consistently, reliably, and at sufficient productivity across different environments.

In plain English: hardware is solved enough to fall to $37,000. Software is not yet solved enough to make every $37,000 unit pay back at 6 months in every environment. The capability ceiling is the gating factor — and the ceiling is being raised by the embodied-AI work happening at Skild, Figure, 1X, Physical Intelligence, Anthropic, and the foundation-model side of the industry, not at the hardware vendors.

Which means the high-utilization industrial deployments (Schaeffler, BMW Spartanburg, Toyota, Hyundai’s Atlas commitment) will fall first; the medium-utilization pilots will limp toward payback; and the consumer home robot is still on the wrong side of the bottleneck for another decade.

The Standard Chartered overlap

The reason this report lands hard is that it published on the same day Standard Chartered attached its own number — 7,000+ back-office positions by 2030 — to the AI-replacement thesis. The IDTechEx forecast is the manufacturing-floor mirror image of the bank announcement.

Both forecasts share the same underlying mechanism: capital costs are falling, capability is rising, and the cost-per-productive-hour line is now crossing the human-cost line in specific scenarios, with the crossing point moving forward year by year. The bank version of this argument has a CEO quote (“job role reductions in favor of the machines”). The manufacturing version of this argument has an IDTechEx cost curve. They are the same argument.

What that means for the labour-market read is that the AI-jobs and robotics stories that this site has been treating as adjacent are converging into a single story — the cost-per-productive-hour line. Banking’s version of the line is risk-reporting and compliance AI. Manufacturing’s version is the humanoid on the box-handling station. The arithmetic is identical.

What to watch

The first humanoid maker to publish ASP under $50,000. IDTechEx’s 2030 forecast is $37,000; Unitree’s G1 was already at ~$16,000 in early 2026 and selling thousands of units. The question is when the industrial-grade humanoids (Figure, Apptronik, 1X, Agility) commit to a public sub-$50k MSRP. That commitment is the canary.
The first publicly disclosed 6-month payback case. IDTechEx models it; somebody has to be the first OEM to file an investor-day slide claiming it. Schaeffler’s Herzogenaurach pilot starting December 2026 is the leading candidate, with BMW Spartanburg’s Figure 03 deployment behind it.
The next major bank or insurer to attach a 2030 number. HSBC’s signalled 20,000 will likely be the next white-collar version. The matched-pair pattern will be: every cost-curve report from a hardware analyst will be followed within weeks by a named headcount cut at a service-sector employer.
The capability ceiling. Whether the humanoid foundation models close the gap to “consistent productive output in unstructured environments” between 2026 and 2030 is the question that decides whether the IDTechEx curve hits its 2036 number or undershoots by half a decade.