Humanoid Robot Forecasts Differ 10x: Who Should You Trust?

These are not competing measurements of the same market. They are different markets, different time horizons, and different assumptions about what counts as a "deployment." If you are timing a home purchase — or just trying to read the headlines honestly — here is how to score each forecaster.

Which Forecast Should Home Buyers Trust?

• Base case for planning: Interact Analysis ~$15B / 700k+ shipments by 2035, with commercial inflection only after 2032.
• Acceleration case: Goldman Sachs central ~$38B / 1.4M units by 2035, requiring faster cost reduction and broader adoption.
• Upside only: Goldman blue-sky ~$154B, which Goldman itself conditions on resolving every major barrier — product design, use cases, technology, affordability, and public acceptance.
• Ignore for home buying: Any 2050 ecosystem number (such as Morgan Stanley's $5 trillion figure covering hardware, software, services, and economic value). It is not a robot-shipment forecast.

Why Do Humanoid Forecasts Differ by 10x?

The Physical AI Journal's September 2026 synthesis of these forecasts makes the key point well: the $15B-to-$154B range is an assumption dispute, not a measurement error.

Interact Analysis counts commercial humanoid hardware shipments into real industrial and commercial applications, grounded in deployment data. Goldman Sachs' central case assumes meaningfully faster cost decline and adoption. The blue-sky case assumes everything goes right.

Three assumption gaps drive almost all of the difference:

1. What counts as deployed. Interact Analysis found only about 10% of humanoids produced were in real-world industrial operations in 2025 — the rest went to research, demos, education, and non-commercial uses. Forecasts that count every unit produced look far more bullish than forecasts that count only commercial deployments.
2. When costs fall. Every aggressive forecast assumes humanoid prices drop fast enough for broad adoption before 2032. Interact Analysis puts the commercial inflection after 2032. That timing gap compounds enormously by 2035.
3. China's role. China plus the US are projected to account for over 85% of 2035 demand, with China alone over 65% of real-world application shipments. But Chinese government procurement — at least $230 million in H1 2026 alone versus $62 million in all of 2025, per Reuters — is policy-driven, not commercially justified industrial demand. Extrapolating global demand from subsidised state purchases overstates the commercial market.

Which Forecaster Has the Best Track Record?

Interact Analysis (May 2026): the most grounded base case

Projection: ~$15B revenue, 700k+ shipments by 2035, inflection after 2032.

Why it earns trust: it is the most recent major benchmark, it separates commercial deployments from research/demo units, and it names the uncomfortable fact — current growth runs on pilots, subsidies, and partnerships, not broad industrial adoption. Its 10%-in-real-operations finding matches what ui44 tracks in shipment data: lots of announcements, few verified commercial deployments. For a buyer, that discipline is exactly what you want in a base case: it tells you what has to go right before timelines compress, rather than assuming the compression and back-filling the justification.

Weakness: conservative by construction. If costs fall faster than expected, it will undershoot.

Goldman Sachs: useful central case, misused blue-sky

Projection: $38B / 1.4M units central (February 2024 analysis, still widely cited); $154B blue-sky; earlier work pointed to ~890k units by 2030 and ~6.5M by 2035 in upside paths.

Why it earns partial trust: the central case is a credible acceleration scenario with explicit conditions. It is the right number to use as a stress test.

Why the blue-sky number misleads: it is routinely quoted without Goldman's own caveat — that it requires resolving every major barrier. A procurement model or purchase-timing decision built on $154B as a base case has misread the source.

Barclays: best near-term deployment tracker

Estimate: ~60,000 deployed in 2026 versus ~15,000 in 2025, scaling toward ~13M units per year by 2035 in aggressive paths.

Why it is useful: near-term deployment counts are checkable, and the 4x year-over-year growth roughly matches the pilot expansion ui44 sees from Figure, Apptronik, Unitree, Agibot, and 1X. The long-dated 13M figure, however, belongs in upside-optionality models, not buying plans.

What Matters More Than Any 2035 Forecast?

BMW's Figure 02 deployment at Spartanburg: 10 months of operation, 10-hour shifts Monday through Friday, 90,000+ components moved, ~1,250 operating hours, supporting production of 30,000+ BMW X3 vehicles.

That is confirmed operation. What BMW has not disclosed: total cost, labour savings, payback period, or comparative economics versus conventional automation. No major platform has published a full public ROI dataset from an independent industrial customer.

For home buyers, the translation is blunt: if industrial customers with engineering teams and measured workflows cannot yet show payback math, home deployment timelines measured in months — not years — deserve heavy skepticism. Vendors announcing home pilots this year are demonstrating teleoperation rigs, choreographed demos, or supervised trials, not autonomous value in an unstructured living room. The BMW data point stays the gold standard precisely because it is boring: shifts worked, parts moved, hours logged — and still no public ROI. Demand that same boredom from any home-humanoid timeline before you believe it.

When Should You Actually Buy a Home Humanoid?

1. Do not time a purchase around the biggest number. The $154B and trillion-dollar figures describe ecosystems and upside cases, not the year a capable home humanoid ships at a price you can afford.
2. Watch the 10% figure, not the headline. When the share of produced humanoids in real commercial operation rises materially above ~10%, forecasts start deserving upward revision. Until then, treat aggressive shipment curves as marketing inputs.
3. Discount China-unit headlines for home buying. High Chinese production volumes plus the July 2026 US import restrictions on Chinese humanoids and quadrupeds (with Unitree among those likely affected) mean units built do not equal units buyable in Western homes.
4. The credible window has not moved. Interact Analysis' post-2032 commercial inflection remains the most defensible planning assumption. Anyone selling you a general-purpose home humanoid "next year" is selling against every grounded forecast in this comparison.

How Should You Read the Next Forecast Headline?

When the next "$X trillion by 2040" headline lands, run this five-question check before changing any buying plan:

1. Revenue or units? Revenue figures can grow through software, services, and fleet contracts while unit counts stay modest. A $50 billion revenue forecast can coexist with fewer than a million robots shipped. Units are what determine whether a home model exists at a price you can pay.
2. Commercial or produced? Ask whether the number counts robots in paid operation or robots rolling off a line into warehouses, labs, and demo floors. The Interact Analysis 10%-in-real-operation finding is the reason headline unit counts overstate the market that matters.
3. Which year does the inflection require? Every bullish curve has a hidden kink year when costs fall and adoption accelerates. If that kink sits before 2030 with no named cost mechanism — no actuator price curve, no battery or compute milestone — treat the curve as aspiration.
4. Who pays in the early years? Government procurement, pilot partnerships, and venture-subsidised deployments do not convert automatically into consumer demand. China's $230 million H1 2026 state buying spree is real spending, but it buys policy outcomes, not proof that homes want humanoids at current prices.
5. What would prove it wrong? A forecast without a falsifiable near-term marker — deployment counts, cost milestones, published ROI — is marketing. The credible forecasts in this comparison all name checkable intermediate numbers. Hold every new forecast to that standard.

Applied to today's market, this checklist keeps pointing at the same conclusion: track verified commercial deployments with published economics, not announcement counts. ui44's robot database exists for exactly that — specs, prices, and shipment signals you can verify instead of headlines you have to trust.

Bottom line: plan on Interact, stress-test on Goldman

Use Interact Analysis as your base case, Goldman central as your acceleration case, and everything above that as upside fiction until deployment economics — not deployment announcements — prove otherwise. The forecasters disagree by 10x because they are answering different questions. Home buyers should only act on one of them: when do verified commercial deployments, with published economics, arrive at scale? On current evidence, that answer is still measured in years, not product cycles.

_Track verified shipments and specs in the ui44 Home Robot Database and compare contenders side by side on our compare page._