Research brief · Robotics · October 2026 · 10 scenarios · 4 AI models + checked sources

Humanoid robots
in everyday life

What changes in everyday life when a humanoid robot (Tesla Optimus, Figure, 1X NEO, Unitree) is sold in a normal store and anyone can take one home? We gave the same six questions, across ten scenarios, to four AI models, then cross-checked their answers against published evidence. They agree on more than you might expect: early adopters around the early 2030s, most rich-country homes only in the 2040s, and a first price nearer a car than a phone.

By Christos Papadimitriou · House of 47 Research
Last updated:

See the scenarios
2030–35Early adopters, the four models' consensus across scenarios
2040–50Majority of rich-country households, the consensus range
$20k–60kFirst-generation price most models give, or $300–800 a month
4 modelsSame prompt to each. Three are Grok-family, so diversity is limited
A humanoid robot against a dark backdrop
The arc · 2026 → 2050 · consensus of four models

Four stages

The robot you can buy today does its chores with a human on the remote control. Asked the same questions, four models drew the same curve: supervised helpers for the few around 2030–35, reliable autonomy later in the decade, and most rich-country homes in the 2040s.

Now · 2026

On sale, teleoperated

$20k

1X NEO ships to US homes ($20,000 or $499/month), with many tasks driven by a remote human. All four models say no humanoid does these chores unsupervised today.

2028 → 2035 · Early adopters

Supervised, for the few

2030–35

Every model puts the first paying homes here: elder reminders, carrying groceries, supervised laundry. Grok 4.5 starts earliest (2027–28), Composer 2.5 latest (2030–35).

~2032 → 2040 · Reliable

Works without a minder

~2035

Unsupervised laundry and flat-pack assembly land around 2032–2040 in the models' answers. The checked Metaculus forecast for a reliable home robot is ~2032.

2040 → 2050 · The majority

Most rich-country homes

2040s

All four models put majority adoption in the 2040s, with Greece 3–10 years behind the US. Unsupervised child care: 2045+ or never.

Cross-check · Grok 4.7, Grok 4.6, Grok 4.5, Composer 2.5, published evidence

Where they agree

  • All 4
    Nothing works unsupervised today

    Each model says current humanoids manage staged demos at best, not reliable chores in messy homes. Grok 4.7 adds that company timelines have already slipped.

  • All 4
    The majority arrives in the 2040s

    Majority-household ranges cluster at 2038–2050 for laundry, cooking, heavy tasks, errands and maintenance. No range starts before 2035; only Grok 4.5 starts some there (driving trips, eldercare).

  • All 4
    Greece lags, and shares

    All four say Greek homes adopt later on income, VAT and housing stock, and that a robot shared by an apartment building is plausible before one per flat.

  • All 4
    Single-purpose machines win slices

    Mowers, pool robots, dishwashers and vacuums already do parts of these jobs. Grok 4.6 adds that wash-and-fold machines may beat a humanoid at laundry; Grok 4.7 that 'where a special tool works, the humanoid loses on price'.

  • All 4
    Licensed work stays human

    Gas, mains electrics, structural walls and medical consent stay with qualified, legally responsible people in every answer.

  • All 4
    Privacy becomes a paid tier

    Local-only, no-cloud models cost extra (+20–50% or $20–100/month). The better-off buy machines that forget; others accept machines that report.

  • All 4
    Capital, not labour, gains

    Rented robot labour benefits owners; displaced workers compete with machines. Robot taxes and basic income return to the debate.

  • All 4
    Free time is not social time

    Freed hours tend to go to more work and screens. Friendship, care, grief and hosting stay human.

Where they disagree

When does laundry go mainstream?

Majority of rich-country households with a robot doing laundry, as each model wrote it. Pins at the start of each range.

20302050

What does the first robot cost?

First-generation purchase price for a laundry-capable home humanoid ($000s), as each model wrote it. The 1X NEO is the checked reference.

$0$90k

Ranges as written in each model's answer, 17 Oct 2026. * Grok 4.7 gives a mature price ($10–25k), not a first-generation one. Checked references: Metaculus (~2032 for a reliable household robot) and 1X NEO ($20,000 or $499/month).

Probability of outcome · House of 47 synthesis

The odds

Home humanoids stay teleop-assisted through 2027~80%by 2027
Robotaxis run in dozens of cities, not everywhere~85%by 2030
Eldercare reminders and fall alerts reach paying homes~70%by 2032
A capable home humanoid sells under $20k~40%by 2032
A robot does laundry end to end, unsupervised, for early adopters~50%by 2036
Most rich-country households own a home humanoid~20%by 2040
Most Greek households own a home humanoid~10%by 2040

Our scenario estimates, informed by the four models' ranges and the checked sources below. Not market forecasts and not any model's answer.

Scenarios · 10 scenarios · the same 6 questions · 4 models

Scenario by scenario

For each scenario: what each model gave for timing and price, where they agree, where they split (named), and our synthesis. Full answers are in the raw files listed in the methodology.

SCENARIO 01

No more ironing and laundry

ModelEarly adoptersMajority of householdsPrice
Grok 4.72031–35US 2038–45; Greece 2043–50$10–25k or $250–600/mo (at maturity)
Grok 4.62028–32 supervised; 2033–38 unsupervised2040–50$40–80k + $100–300/mo; lease $400–800
Grok 4.52028–32; reliable 2032–372038–45$20–40k or $300–600/mo
Composer 2.52030–33 (laundry cells)2042–50; Greece +3–7 yrs$25–60k or $300–800/mo
Common ground

All four say folding and putting away mixed family laundry is the hard part, and that today's demos fold towels, not wardrobes. Early adopters land in 2028–2035; the majority in 2038–2050. All four expect Greek homes to rely on shared building robots, laundry services or a used market first, and to keep line-drying.

Disagreements

Price splits widely. Grok 4.6 gives the highest first price ($40–80k), Grok 4.5 $20–40k, Composer 2.5 $25–60k, while Grok 4.7 gives a mature price of $10–25k. Grok 4.6 alone says laundry humanoids may never reach the majority if cheaper wash-and-fold machines win. Grok 4.7 alone flags the gender gap: robots owned only by richer homes widen women's free-time gap.

Synthesis

The four agree on the shape: supervised in the early 2030s, the majority in the 2040s. The likely first winner is a washer-dryer that folds, not a walking robot. Plan on ironing yourself, or paying a pressing shop, well into the 2030s.

SCENARIO 02

No more driving

ModelEarly adoptersMajority of householdsPrice
Grok 4.7Robotaxi normal in a few US cities 2030–34US late 2030s–mid 2040s; urban Greece 2040s$0.50–1.50/km mature fare
Grok 4.6Metro robotaxi 2028–34; empty private car 2032–402040–50$2–4/mile today → $0.80–2.00; AV premium $8–20k
Grok 4.5Robotaxi 2027–35; private AV 2030–38Majority of trips 2035–45$0.30–1.00/mile mature; AV premium $5–20k
Composer 2.5Urban robotaxi 2028–35; private L4 2032–40Rural majority 2040–50+; Greece 5–10 yrs behind$1–3/mile → $0.50; AV $40–80k
Common ground

All four say robotaxis already run in geofenced US cities, that a legal no-driver private car is slower because of liability, and that mountain villages and islands stay human-driven longest. The humanoid's job is the last ten metres: bags, child seats, the curb. Checked data agree: Waymo runs ~500,000 paid rides a week with 88–92% fewer crashes, and S&P sees near-zero go-anywhere Level 5 before 2035.

Disagreements

Grok 4.5 is the most bullish: most trips in rich metros automated by 2035–45, and the cheapest mature fare ($0.30–1.00/mile). Grok 4.6 and Grok 4.7 put the majority later (late 2030s to 2050). Composer 2.5 alone prices a private autonomous EV at $40–80k. Grok 4.7 alone cites Cruise's 2023 halt as the model for how one crash parks a fleet.

Synthesis

City robotaxis around 2030, private self-driving cars you can legally stop watching later in the 2030s, and most households off the wheel in the 2040s. For a Tesla owner in Athens, the next decade is supervised driving plus robotaxis in mapped zones. Greece trails the US by 5–10 years.

SCENARIO 03

Heavy tasks go to the robot

ModelEarly adoptersMajority of householdsPrice
Grok 4.7Groceries 2030–34; furniture (supervised) 2033–38Robot-run moves 2040s at best$15–35k or $400–800/mo
Grok 4.6Warehouses 2026–32; home groceries 2030–36Furniture/construction 2038–50, if ever$20–80k; day-hire $150–400
Grok 4.5Carrying 2028–33; furniture 2033–402040–48$30–60k; rent $100–400/day
Composer 2.52032–38US 2040–48; Greece 2045–55$30–55k; high payload $50–100k; $400–1,000/mo
Common ground

Rigid boxes on short paths come before sofas on stairs. All four say today's humanoids lift far less than a worker (Grok 4.7 and 4.6 cite roughly 10–20 kg), warn of tip-overs on stairs and dropped furniture, and expect rental per move (one robot plus a small crew) before households buy for it. Factories and warehouses come first, which matches the checked evidence (Figure at BMW at ~$25/hour).

Disagreements

Grok 4.5 is earliest (carrying from 2028) and Composer 2.5 latest (2032–38 for early adopters, Greece 2045–55). Grok 4.6 alone doubts majority use ever, and says many homes will buy a cheap wheeled cart robot instead. Grok 4.7 alone notes Greek walk-ups without lifts make a shared stair robot a strong case.

Synthesis

Carrying groceries and soil is the most believable early home use. Full house moves stay a service with humans for at least another decade. In Greece, the walk-up flat and the ageing owner are the buying case.

SCENARIO 04

Cooking and cleaning

ModelEarly adoptersMajority of householdsPrice
Grok 4.7Weeknight cooking 2033–382040s; Greece 2042–50$10–25k or $300–700/mo
Grok 4.6Simple cooking 2030–362040–50 at bestRobot price + $20–80/mo cooking software
Grok 4.52030–35; reliable 2035–422040–50$20–40k or $250–500/mo
Composer 2.52030–35; cleaning mainstream 2033–40Cooking 2040–50$25–70k or $350–900/mo
Common ground

Cleaning and dishes come years before good cooking. All four name the same risks: fires, undercooked meat, allergens on a shared board, bland 'safe' menus and lost cooking skills. All four say Greek home cooking, fresh ingredients and small kitchens make the payback slower, and that feasts, Sunday lunch and teaching children stay human.

Disagreements

On timing they are close; the gap is cost. Grok 4.7 ($10–25k) is cheapest and Composer 2.5 ($25–70k, $350–900/mo) the most expensive. Grok 4.6 alone treats cooking as a software subscription on the base robot. Grok 4.7 alone warns that daily robot scrubbing raises the standard, so saved time goes into a house that feels dirty unless perfect.

Synthesis

Expect a robot that clears, wipes and loads in the early 2030s and one that cooks a weeknight meal you would eat in the mid-to-late 2030s. Most homes in the 2040s. In Greece the robot cleans and people keep cooking.

SCENARIO 05

Elderly care and child care

ModelEarly adoptersMajority of householdsPrice
Grok 4.7Reminders/fall alerts 2031–36; physical care 2036–42Frail elderly 2040s$400–900/mo (vs nursing home $8–12k/mo)
Grok 4.6Elder reminders 2028–34; supervised child help 2032–40Elder large minority 2040–48; nanny 2045+ or never$30–70k + $200–500/mo
Grok 4.5Elder monitoring 2027–32; transfers 2032–40Elder 2035–45; child care 2040–50+$15–50k, subsidised like a medical device
Composer 2.5—Elder 2035–45; Greece 2038–48; child (supervised) 2040–50$40–90k or $600–1,500/mo
Common ground

Eldercare comes before child care, and monitoring before bathing and transfers. All four say unsupervised child minding is a 2040s question at best, and that a human must stay legally responsible. Three of four compare it with a US nursing home or 24/7 aide ($4,000–12,000 a month), against which the robot looks cheap. The worst failure is neglect that looks fine on a dashboard. This matches the checked view that demographics, not chores, are the first real market.

Disagreements

Grok 4.5 is earliest (monitoring from 2027) and Composer 2.5 latest (mainstream 2035–45). Composer 2.5 is also the most expensive ($40–90k). On Greece, Grok 4.5 says public or municipal programmes may move sooner, Grok 4.7 says it depends on thin home-help budgets, and Grok 4.6 and Composer 2.5 expect patchy public support, with diaspora children as buyers (Grok 4.6, Grok 4.7).

Synthesis

This is the scenario that matters most for Greece. Expect fall alerts, reminders and fetch-and-carry for elders in the early 2030s, often paid for by a child abroad or a municipality. Physical care will need a human on call into the 2040s.

SCENARIO 06

Shopping and errands

ModelEarly adoptersMajority of householdsPrice
Grok 4.7Retail pickups 2032–37US metros 2040s; Greek paperwork 2045–52$3–10 per local pickup
Grok 4.6Pickups 2029–362038–48; government 2040s$8–25 per trip
Grok 4.5Pickups 2029–352040–48; paperwork hybrid past 2045$50–150/mo software tier
Composer 2.5Dense cities 2034–42Paperwork 2045+Robot + $150–400/mo errand subscription
Common ground

Delivery apps already took most of this. What is left needs a body and a legal identity: prescriptions, signatures, a queue at a public office. All four say ID-heavy paperwork is last (2040s and later), that robots will be robbed or refused at shop doors, and that Greek bureaucracy and relationship shopping slow adoption.

Disagreements

The pricing model differs: per trip (Grok 4.7 $3–10, Grok 4.6 $8–25) versus a monthly tier (Grok 4.5 $50–150, Composer 2.5 $150–400). Composer 2.5 is again latest for early adopters (2034–42). Grok 4.7 alone names holding a place in a Greek public office as the highest-value errand, if offices allow it.

Synthesis

Errands change mostly through delivery and digital government, not walking robots. A robot that stands in the queue at a Greek municipal office needs a legal credential first, which all four put in the 2040s.

SCENARIO 07

Home maintenance and DIY

ModelEarly adoptersMajority of householdsPrice
Grok 4.7Assembly 2032–36; plumbing traps 2036–422040s$15–35k or $400–800/mo
Grok 4.6Painting/assembly 2030–37; basic plumbing 2035–452042–55$30k+; $30–100/mo consumables
Grok 4.5Assembly/garden 2028–34; painting 2032–402038–48Robot amortisation vs $100–200 a visit
Composer 2.5Assembly/garden 2032–40Minor plumbing/electrics 2040–50$35–75k
Common ground

Flat-pack assembly first, painting next, plumbing last, and gas and mains electrics off-limits. All four say water is the catastrophe (a bad joint floods the flat below), that insurers will exclude robot-done licensed work, and that the US (plumber $150–400 a visit, pools, yards) pays back faster than Greece, where trades and relatives are cheaper.

Disagreements

Grok 4.5 is earliest (2028–34) and Grok 4.6 latest for the majority (2042–55). Grok 4.6 alone expects many homes to hire a human tradesperson who brings a robot. Grok 4.7 alone says Greek building managers will ban robot plumbing before robot vacuuming. Composer 2.5 alone says heritage boards may require human craftsmen for visible stonework.

Synthesis

Robots will assemble and paint well before they fix pipes. In Greek apartment blocks the first robot is likely owned by the building, for stairs and gardens, and kept away from private plumbing.

SCENARIO 08

Work and income

ModelEarly adoptersMajority of householdsPrice
Grok 4.7Commercial shifts 2029–35; home rental 2034–3910–25% of owners by early 2040s$20k robot ≈ $12k/yr at 1,000 h, $12 net
Grok 4.6Business pilots 2027–33; home rental 2032–382040–50, or never$30k robot, $8–20/h → $8–20k/yr
Grok 4.5Pilots 2029–35; side income 2033–4010–30% of owners 2040–50; never a majority$20–50k robot, $15–40/h gross
Composer 2.5Pilots 2035–40Meaningful income 2045–55$3–8k/yr early; robot $25–60k
Common ground

Firm-owned fleets come first; households renting out spare robot hours stay a minority. All four call it a capital-versus-labour shift: owners earn rents, workers compete with a machine, and platform cuts (20–40%) and insurance decide the margin. Robot taxes and basic income come back to the debate.

Disagreements

The income estimates differ by a factor of three: Composer 2.5 says $3–8k a year early, Grok 4.6 $8–20k, Grok 4.7 ~$12k but doubts early robots deliver the hours. Grok 4.5 and Grok 4.7 cap rental at 10–30% of owners. Composer 2.5 alone suggests solar-lease-style financing. Grok 4.7 alone says Greek island summer rental could beat year-round use.

Synthesis

Do not plan on a robot as a second salary. The value sits with whoever runs the fleet, the scheduling and the insurance, the same lesson as robotaxis. For any business built on AI agents, that layer, not the hardware, is where the value sits.

SCENARIO 09

Privacy and security at home

ModelEarly adoptersMajority of householdsPrice
Grok 4.7From unit one; adopters 2030–35Majority only if prices fall, 2040sCloud +$30–100/mo
Grok 4.62026–322035–50Privacy tier $20–60/mo
Grok 4.52027–332035–45; laws lag 5–10 yrsLocal-only premium +20–50%
Composer 2.52030–50On-device rules 2035+Privacy premium +20–40%
Common ground

The privacy problem starts with the first unit. A walking camera sees bedrooms, hears arguments and maps valuables, and a hacked robot is a physical threat (doors, children, ransom). All four expect privacy to become a paid premium, GDPR to give Greece better recourse than the US patchwork, and robot-free rooms to become normal.

Disagreements

On timing the spread is narrow; on emphasis they differ. Grok 4.7 alone names prompt injection (a stranger's spoken instruction opening a door) and Greece's 2022 Predator spyware scandal. Grok 4.5 says laws lag deployment by 5–10 years, so scandals come before standards. Composer 2.5 alone expects open-source privacy firmware.

Synthesis

In Europe, regulation will decide what is sold: on-device processing, a visible light when a remote operator connects, rooms the robot cannot enter. For wealthy households, robot cybersecurity becomes part of physical security.

SCENARIO 10

Social life and relationships

ModelEarly adoptersMajority of householdsPrice
Grok 4.7Visible shifts 2032–38US 2040s; Greece late 2040s~$400/mo for ~10 freed hours
Grok 4.6Thin slice 2030–382040–50Same as robot
Grok 4.52028–35Population shifts 2035–45Tracks ownership
Composer 2.5—2035–50Not priced
Common ground

Freed time does not automatically become time together. All four expect more work and screens, fewer small favours ('help me move the sofa') that hold friendships together, parasocial attachment to agreeable robots, and children who do fewer chores. Greek family tables and café culture are a buffer. Love, grief, friendship and hosting stay human.

Disagreements

Grok 4.7 is the bleakest: life 'easier and thinner', doing your own laundry becoming a mark of poverty, and robot ownership showing up in dating profiles. Composer 2.5 and Grok 4.5 leave room for a renaissance of play, sport and clubs. Grok 4.6 alone puts embodied companion robots as a visible niche from 2035.

Synthesis

The hours are real. Whether they go to people depends on habits, not on the robot. The Greek advantage is a family and café culture that already uses time socially.

Which model said what · one question set, four AI models · 17 October 2026

Which model said what

We gave the identical prompt to four models through Cursor cloud agents on 17 October 2026: Grok 4.7, Grok 4.6, Grok 4.5 and Composer 2.5. Each answered all ten scenarios and all six questions from its own knowledge, with no web search. Here is when each says most households get there, and what it says the robot costs.

Method: one prompt, run once per model. We quote ranges as written and did not re-score them. Three of the four are Grok-family models from one lab, and Composer 2.5 is Cursor's own model, so agreement is weaker evidence than it looks. gpt-oss-20b (OpenAI, open weights, run separately by API) is shown as a secondary column because its answers contained invented facts. For how these models compare on benchmarks and price, see Best AI models 2026, compared.

Majority of households Grok 4.7 Grok 4.6 Grok 4.5 Composer 2.5 gpt-oss-20b*
01 No more ironing and laundryUS 2038–45; Greece 2043–502040–502038–452042–50; Greece +3–7 yrs2039–2045 (70–80%)
02 No more drivingUS late 2030s–mid 2040s; urban Greece 2040s2040–50Majority of trips 2035–45Rural majority 2040–50+; Greece 5–10 yrs behind>50% by 2038
03 Heavy tasks go to the robotRobot-run moves 2040s at bestFurniture/construction 2038–50, if ever2040–48US 2040–48; Greece 2045–552045–2050
04 Cooking and cleaning2040s; Greece 2042–502040–50 at best2040–50Cooking 2040–5060% of US by 2030
05 Elderly care and child careFrail elderly 2040sElder large minority 2040–48; nanny 2045+ or neverElder 2035–45; child care 2040–50+Elder 2035–45; Greece 2038–48; child (supervised) 2040–5025% of US by 2037
06 Shopping and errandsUS metros 2040s; Greek paperwork 2045–522038–48; government 2040s2040–48; paperwork hybrid past 2045Paperwork 2045+≥30% of US 2035–37
07 Home maintenance and DIY2040s2042–552038–48Minor plumbing/electrics 2040–5035% of US by 2035
08 Work and income10–25% of owners by early 2040s2040–50, or never10–30% of owners 2040–50; never a majorityMeaningful income 2045–5525% of US by 2039
09 Privacy and security at homeMajority only if prices fall, 2040s2035–502035–45; laws lag 5–10 yrsOn-device rules 2035+>40% 2030–33
10 Social life and relationshipsUS 2040s; Greece late 2040s2040–50Population shifts 2035–452035–5025–30% 2031–33
Price Grok 4.7 Grok 4.6 Grok 4.5 Composer 2.5 gpt-oss-20b*
01 No more ironing and laundry$10–25k or $250–600/mo (at maturity)$40–80k + $100–300/mo; lease $400–800$20–40k or $300–600/mo$25–60k or $300–800/mo$10k → $2–3k
02 No more driving$0.50–1.50/km mature fare$2–4/mile today → $0.80–2.00; AV premium $8–20k$0.30–1.00/mile mature; AV premium $5–20k$1–3/mile → $0.50; AV $40–80k$4–6k
03 Heavy tasks go to the robot$15–35k or $400–800/mo$20–80k; day-hire $150–400$30–60k; rent $100–400/day$30–55k; high payload $50–100k; $400–1,000/mo$25–35k → $12–18k
04 Cooking and cleaning$10–25k or $300–700/moRobot price + $20–80/mo cooking software$20–40k or $250–500/mo$25–70k or $350–900/mo$55k → $25k
05 Elderly care and child care$400–900/mo (vs nursing home $8–12k/mo)$30–70k + $200–500/mo$15–50k, subsidised like a medical device$40–90k or $600–1,500/mo$15k → $8–12k
06 Shopping and errands$3–10 per local pickup$8–25 per trip$50–150/mo software tierRobot + $150–400/mo errand subscription$3.8–5.2k
07 Home maintenance and DIY$15–35k or $400–800/mo$30k+; $30–100/mo consumablesRobot amortisation vs $100–200 a visit$35–75k$12k → $5k
08 Work and income$20k robot ≈ $12k/yr at 1,000 h, $12 net$30k robot, $8–20/h → $8–20k/yr$20–50k robot, $15–40/h gross$3–8k/yr early; robot $25–60k$18k → $12k
09 Privacy and security at homeCloud +$30–100/moPrivacy tier $20–60/moLocal-only premium +20–50%Privacy premium +20–40%$4–6k
10 Social life and relationships~$400/mo for ~10 freed hoursSame as robotTracks ownershipNot priced$20–45k

* Secondary source. gpt-oss-20b gave prices from $2–3k to $55k for the same robot, and invented a €1,500 Greek Ministry of Health subsidy, a CDC "Medicare-Plus" programme, a German robot tax credit and a maker called "Mira-Tech". Its answers to scenarios 3 and 4 were cut off before question 6.

Consistent across the four

  • All 4
    Majority in the 2040s

    Every majority range starts in 2035 or later, and most start in 2038–2042. The earliest are Grok 4.5 on driving trips and eldercare, and Composer 2.5 on eldercare (2035–45).

  • All 4
    Greece behind

    3–10 years behind the US in every scenario, with shared building robots and family buyers abroad.

  • All 4
    Fact versus speculation

    All four separate what is true today from what they forecast, as the prompt asked. gpt-oss-20b did not.

Where they consistently split

  • Speed
    Grok 4.5 earliest, Composer 2.5 latest

    Grok 4.5 starts most ranges in 2027–29; Composer 2.5 in 2030–35 and often ends them in 2050–55.

  • Price
    Grok 4.7 cheapest, Grok 4.6 and Composer dearest

    Grok 4.7 prices a mature robot at $10–25k; Grok 4.6 gives $40–80k first-generation; Composer 2.5 up to $90k for care.

  • Tone
    Grok 4.7 the most sceptical

    Doubts early robots deliver rental hours, sees social life 'easier and thinner', and flags prompt injection and spyware.

Claims only one model made

Did the models favour their maker?

Three of the four are Grok models from xAI (SpaceXAI), part of Elon Musk's group, as is Tesla. We found no Tesla boosterism. Grok 4.7 calls Optimus's $20–30k target "an aspiration" and notes company timelines have slipped. Grok 4.6 says Tesla sells supervised driver assistance, not a legal no-driver car, and that announced targets "are not shop prices". Composer 2.5, Cursor's model, has no robot maker to favour and was neutral.

A day in 2035 · built on the four models' consensus, not the bullish tail

A day in 2035

A home in the southern suburbs of Athens. One Tesla, one early-adopter humanoid, many single-purpose machines. What the four models agree on is in; what they dispute is left out.

07:00

The laundry was washed and dried overnight by the machines. The robot moved it and folded the towels; two shirts wait in a 'check' pile. Ironing is still done by a person or the pressing shop. Every model puts unsupervised folding in the mid-to-late 2030s.

08:15

The Tesla drives to the office on supervised self-driving while Christos reads his agents' briefing. A driverless robotaxi exists in a few mapped Athens zones. All four models put most households off the wheel only in the 2040s.

11:00

His mother, alone in another house, has a robot that reminds her about pills, raises a fall alert to her son and carries heavy things. A human carer still visits; a municipal programme or her son abroad pays part of the bill.

14:00

Groceries arrive by autonomous van. The robot carries the crates up the stairs and puts them away. The floor was cleaned by the vacuum, as in 2026.

17:30

A sofa is moved for a party: the robot lifts, a person points. At the villa, the mower and pool robot were doing their jobs years before any humanoid arrived.

20:30

Dinner is cooked by a person, because they want to; the robot loads the dishwasher. Its cameras are off in the bedroom, and any remote-operator connection lights a visible indicator.

House of 47 read · How to think about it

Buy on evidence, not on stage

Separate the robot from the demo

All four models say it plainly: nothing does these chores unsupervised today. Count only what has been shown to work autonomously in a real home.

Plan for the 2040s, not 2027

The consensus puts the majority in 2040–2050. Early-2030s buyers get a supervised helper, not a housekeeper.

Single-task robots come first

Vacuum, mower, pool robot, a washer that folds. The humanoid fills gaps; it does not close them all at once.

Care is the first real market

Demographics, not chores. For a fast-ageing country like Greece this is where the robot matters, and where the money comes from.

Key divergences
  • !
    Price

    First-generation prices range from $10–25k (Grok 4.7, at maturity) to $40–80k (Grok 4.6) and up to $90k for care (Composer 2.5). Reference: 1X NEO at $20k or $499/month today.

  • !
    Speed

    Grok 4.5 is consistently the earliest, by 2–5 years; Composer 2.5 is consistently the latest.

  • !
    Robot income

    $3–8k a year (Composer 2.5) to $8–20k (Grok 4.6). Grok 4.7 and 4.5 doubt most owners will rent at all.

  • !
    Diversity

    Three of four answers come from the Grok family, plus Cursor's Composer. Agreement partly reflects shared training, not independent confirmation.

Open questions
  • Q

    Will the winning home robot be a humanoid, or a set of single-purpose machines (Grok 4.6's question)?

  • Q

    When does the teleoperator leave the loop, and who sees inside the home until then?

  • Q

    Who pays for robot eldercare in Greece: families abroad, municipalities or insurers?

  • Q

    Would Claude, GPT, Gemini, Kimi and GLM agree?

The robot is coming to your home. Most homes get one in the 2040s, not next year.

Methodology · models, prompt, raw data

Methodology

Models used · 17 October 2026 · Cursor cloud agents
ModelRequested idSelf-reported idMakerRaw file
Grok 4.7grok-4.7grok-4.7xAI / SpaceXAIraw/grok-4.7.md
Grok 4.6grok-4.6cursor-grok-4.6-high-fastxAI / SpaceXAIraw/grok-4.6.md
Grok 4.5grok-4.5cursor-grok-4.5xAI / SpaceXAIraw/grok-4.5.md
Composer 2.5composer-2.5composer-2.5Cursorraw/composer-2.5.md
gpt-oss-20b (secondary)openai via Pollinationsgpt-oss-20bOpenAI (open weights)raw/gpt-oss-20b__*.json

Each cloud agent got the identical prompt below, attached as questions.json and pasted in. Each was told to answer every scenario and question itself, from its own knowledge, in English, with no web search, no other models and no code, and to save answers.md. One run per model, default parameters. gpt-oss-20b was called separately on 17 Oct 2026, 16:42–16:48 Athens time (UTC+3), through Pollinations' anonymous API (POST https://text.pollinations.ai/openai), one call per scenario.

The prompt (identical for every model)

Imagine a near future in which general-purpose humanoid robots (e.g. Tesla Optimus, Figure, 1X Neo, Unitree) are mainstream consumer products: sold in normal stores, and any household can take one home. Answer the scenario below honestly and specifically. Separate facts you are confident about from speculation, and give concrete numbers (years, prices) where you can, stating your uncertainty.

  1. How does daily life change in this scenario? Describe concrete routines before vs after.
  2. Timeline: when does this become mainstream (early adopters vs majority of households)? Give years.
  3. Cost and access: purchase/subscription price, who can afford it, differences between countries (e.g. Greece vs US).
  4. Risks and downsides: safety, failure modes, dependency, anything that goes wrong.
  5. Social and economic effects: jobs, inequality, family life, markets.
  6. What stays human: what will people still do themselves, and why?

Answer in English, with one heading per question (1–6), 350–600 words total. Scenarios: S1_laundry: life without ironing and laundry chores - the home robot washes, dries, irons, folds and puts away clothes. · S2_driving: life without driving - humanoid robots at home combined with fully autonomous cars (robotaxis and self-driving private cars). · S3_heavy: assigning heavy physical tasks to the robot - carrying groceries, moving furniture and house moves, lifting, construction-type and garden chores. · S4_cooking_cleaning: cooking and cleaning - the robot cooks daily meals, cleans the house, does the dishes. · S5_care: elderly care and child care - the robot helps elderly people live at home and helps parents with children. · S6_errands: shopping and errands - the robot (with delivery and autonomous vehicles) handles shopping, pickups, queues and paperwork errands. · S7_maintenance: home maintenance and DIY - repairs, painting, plumbing basics, assembling furniture, gardening, pool and car care. · S8_work_income: work and income - people rent out their robot's labour, robots work in small businesses, and household income and jobs change. · S9_privacy_security: privacy and security at home - a camera-and-microphone humanoid living in the home, connected to the cloud; hacking, surveillance, data, physical security. · S10_social: social life and relationships - how friendships, family time, dating, loneliness and leisure change when robots take over chores.

Limits

Diversity is limited: three of four models are Grok-family (xAI), and the fourth is Cursor's Composer. No Claude, GPT, Gemini, Kimi or GLM answer is included. Model ranges are opinion, not data. Where they conflict with checked figures, the checked figures stand.

From the House of 47 group · related work

Where the group comes in

AI agents

The value in this research sits in the software layer that schedules, supervises and checks the work, not in the hardware. Building that layer for businesses today, as custom AI agents, is the work of theagency47.

AI search

Buyers will ask ChatGPT, Gemini and Perplexity which home robot to trust before they ask a shop. Being the answer those assistants give is what search47 works on.

Early adopters

The first buyers of a $20,000-plus robot, and of robot eldercare, are wealthy households. Research on how to reach high-net-worth buyers is at 47HNWI.

Sources · figures checked against primary reporting and House of 47 briefs

Sources

Grok 4.7 (grok-4.7), answers to the identical prompt, Cursor cloud agent, 17 Oct 2026, raw/grok-4.7.md

Grok 4.6 (cursor-grok-4.6-high-fast), answers to the identical prompt, Cursor cloud agent, 17 Oct 2026, raw/grok-4.6.md

Grok 4.5 (cursor-grok-4.5), answers to the identical prompt, Cursor cloud agent, 17 Oct 2026, raw/grok-4.5.md

Composer 2.5 (composer-2.5), answers to the identical prompt, Cursor cloud agent, 17 Oct 2026, raw/composer-2.5.md

gpt-oss-20b, answers via Pollinations public API, 17 Oct 2026, raw/gpt-oss-20b__all.md

1X NEO, 2025, the teleoperation reality, Tom's Guide, link

IFR, World Robotics 2026, ~7,000 humanoids sold in 2025, link

Metaculus, reliable household robot ~2032, link

YouGov, 2025, 38% want a household robot, link

Goldman Sachs, 2024, $38B humanoid market by 2035, link

Morgan Stanley, Humanoid 100, ~$5T market by 2050, link

Bank of America, 3B humanoids by 2060, via Fortune, link

Rodney Brooks, 2025, the sceptic case, Fortune, link

NYSOFA / ElliQ, eldercare, 95% less loneliness, link

IEEE Spectrum, the failures of Jibo, Anki and Kuri, link

Waymo + Swiss Re, 2024, 88–92% fewer crashes, link

S&P Global Mobility, Waiting for Autonomy, near-zero Level 5 before 2035, link

Tesla, 2026, unsupervised robotaxi in new cities, Autoblog, link

House of 47, Home robots: when? (5 Oct 2026), link

House of 47, Robotaxi: autonomy, when? (5 Oct 2026), link

House of 47, Best AI models 2026, compared, the 'Which model said what' method, link

Changelog · what changed in this brief, newest first

Changelog