What this post is and isn’t

This is a qualitative state-of-the-market note. We do not quote unit-shipped figures because public robot numbers for 2026 are either estimates or vendor marketing, and we would rather say “we observe” than invent statistics. If you want precision, IFR (International Federation of Robotics) publishes their annual report in autumn and is the source of record.

Hardware side

Consumer humanoids entered the £15k-£35k bracket during 2025-2026 — within reach of high-income households but not mass market. Quadrupeds are cheaper (£2-5k) and more widely deployed as security and inspection devices. The installed base for skill-compatible consumer robots remains small in absolute terms; what is striking is the slope of the curve, not the magnitude.

Developer side

Developer supply is ahead of buyer demand — the familiar shape of a new marketplace. We see this on GeraSkills: submission volume outpaces install volume by roughly the same ratio the iOS App Store showed in 2009. This is a normal phase; it reverses as hardware penetration climbs.

Category distribution

The top category clusters on our platform are: domestic (laundry, cleaning, cooking assistance), pet care (feeding, walking, play), health-adjacent (reminders, mobility assistance), and hobby (gardening, model-making assistance). The two under- served categories are eldercare (regulated, high-bar-of-evidence) and child-interaction (safety concerns, limited willingness-to-publish).

Regulatory pressure

The EU AI Act classifies several robot-skill categories as high-risk — anything involving physical contact with vulnerable persons triggers extra obligations. The UK is still catching up; the US is state-by-state. GeraSkills complies via GeraCompliance and our Trust Centre publishes the exact obligation map per skill category.

What creators should take away

Supply is easy; demand will catch up. Build skills for the 2027 market, not the 2026 one.
Compete on polish and trust, not novelty. The existing laundry skills all work; the one that wins has the best edge-case handling.
Avoid the under-served categories unless you have domain credentials. Eldercare is lucrative; eldercare without credentials is a lawsuit.

What buyers should take away

Check signature verification badges. Every legit skill is signed.
Read permission lists. A laundry skill that requests kitchen access is a red flag.
Try the free ones first. The 70/30 split means creators want you to try before you buy; many offer free tiers.

What we will be watching into 2027

Hardware price trajectory, eldercare regulation clarity, the first cross-platform skill standard (candidates exist; none has won), and the eventually-inevitable high-profile safety incident that shapes public opinion. We will write the 2027 review against these four trendlines.

Cross-links

See also: 2027 predictions, best-selling categories, the coming economy. Regulatory context via GeraCompliance; payments via GeraCash.