7The Next Decade

The market in 2031 and 2036

The argument so far is structural. But if coordination gets cheaper, homes become better understood, and machines begin performing physical work, what happens to the size of the market?

A bottom-up model suggests something more interesting than ordinary industry growth.

Start with the homes

The U.S. had 148.3 million housing units in 2025, including the 82.9 million owner-occupied homes of Section 1. [3][16] Harvard estimates that owners spent $105 billion on routine maintenance in 2023, while 49% of $405 billion in improvement spending went to need-to-do replacements such as roofing, windows, plumbing, electrical, and HVAC. [17]

After adding rental properties and recurring services, while excluding discretionary remodeling and new construction, this paper estimates the 2026 core home-services economy at roughly $400 billion, or about $2,700 per U.S. housing unit: the recurring, need-driven core of the more than $600 billion in total spending described in Section 1.

This is an analytical baseline, not a reported TAM. The model then applies five forces:

The first three are conventional. The last two are where the market changes shape.

The maintenance-need assumption reflects a documented trend: the U.S. housing stock is the oldest on record, with owner-occupied homes at a median age of roughly 42 years. [17]

AI reduces the effort required to buy a service. More jobs that are postponed, bundled into a future weekend, or simply ignored become transactions. Robotics can move the other constraint: the cost and availability of doing the work.

A path toward $1 trillion

Nominal dollars. Scenario estimates are the author's analysis, not third-party forecasts. Robot hardware sales are excluded.

The important jump is not inflation. Without meaningful robotics, housing growth, aging systems, price increases, and easier coordination take the market toward roughly $700–750 billion by 2036. Probots can push it closer to $1 trillion because they do more than replace labor.

A probot, as used here, is a professional robot that performs or materially assists paid physical work at a home.

Cleaning, lawn care, inspection, gutter work, exterior maintenance, equipment diagnostics, material handling, and standardized repairs are likely to automate before complex plumbing, electrical, or HVAC work. Professional service robotics is already an established category: nearly 200,000 professional service robots were sold globally in 2024, with rapid growth in cleaning and inspection applications. [18]

The economic effect comes through four channels:

More capacity. Machine hours supplement scarce professional hours.

Smaller jobs become viable. A loose hinge or minor maintenance task may not justify a human truck roll. A machine already servicing the home can do several such tasks at once.

Service becomes more frequent. Lower execution costs make preventive maintenance more attractive than waiting for failure.

More problems are discovered. Home records, sensors, agents, and machine inspection identify work homeowners would otherwise never request.

The result is unusual: the price of some individual tasks can fall while the total market grows. A home that currently buys four services a year might buy six, eight, or ten when finding, scheduling, and performing small jobs becomes inexpensive. That is how physical automation expands the market rather than merely dividing today's labor bill between humans and machines.

The coordination economy grows with it

Robotics does not remove the coordination layer. It gives it more to coordinate. Today the system chooses among professionals. Tomorrow it chooses among professionals, specialist robots, humanoids, remote operators, and combinations of all four.

If software orchestrates roughly 20% of service GMV by 2031, an 8–12% blended economic capture implies a $9–13 billion orchestration revenue pool. If roughly half of a ~$1 trillion market is orchestrated by 2036, the same economics imply roughly $40–60 billion of annual revenue flowing to the coordination layer.

That revenue is not additional to service GMV. It is the portion captured by systems that price, route, schedule, pay, guarantee, and remember the work. And as execution becomes more abundant, the scarce asset shifts upstream.

The home record provides context. The orchestration layer makes the decision. Humans and machines provide execution.

Robots change who holds the screwdriver, not who holds the judgment.

2030 base case

In major metros, AI-mediated intake, scheduling, status, documentation, and follow-up are normal, and photo and video triage is standard. Standardized jobs carry guaranteed prices; complex jobs carry bounded ranges or paid diagnostics. The professional is still human; far less human labor sits between homeowner and technician. Adoption is uneven by design: cleaning, landscaping, pest control, and routine maintenance first; hidden-condition plumbing, electrical, and major HVAC later; rural markets lag because software cannot manufacture local supply.

2036 expected direction

By 2036, the home record becomes increasingly persistent and portable. Equipment, warranties, past work, and verified outcomes travel across providers and ownership changes.

Fulfillment becomes mixed: professionals, professionals using robots, and probots performing standardized work under supervision. Complex and safety-critical trades remain more human-led.

This makes orchestration more important, not less. The system must decide whether each job goes to a human, a machine, or both.

Robots change who holds the screwdriver, not who holds the judgment.

The robust prediction is not that robots fix every home. It is that far less human effort is spent arranging, and increasingly performing, routine work.

Five predictions

1. By 2030, no standalone lead-generation model ranks among the three most valuable companies in home services. Value concentrates in repeat relationships and completed-job economics.

2. By 2030, in the top-20 U.S. metros, a majority of routine jobs (cleaning, landscaping, recurring maintenance) on managed platforms are booked, scheduled, and paid end-to-end with no human coordinator.

3. By 2032, at least one AI-native coordinator exceeds $1B in annual completed-job GMV with over 60% repeat or embedded demand.

4. By 2036, portable home records cover at least 20% of U.S. owner-occupied homes, with at least one insurer or manufacturer subsidizing them at scale.

5The 2036 category leader is not a 2026 consumer marketplace. It emerges from orchestration, pro-side software, or a new entrant, and execution economics decide which, not chat experience.

Where this thesis breaks

• General assistants capture the consumer interface and vertical platforms fail to build defensible assets behind it.
• Quote-before-a-visit variance stays too high to underwrite profitably outside narrow standardized categories.
• The best professionals reject platforms, forcing take rates down and leaving networks with adverse-selection supply.
• Worker-classification, licensing, or insurance rules in key states make managed orchestration hard to run profitably.
• Home data stays fragmented because consumers, manufacturers, contractors, and insurers never permit a portable record.

None of these risks is hypothetical; all are visible today. None has yet proven fatal to the shift toward software coordination.

Home services will not be won by the company that answers first. It will be won by the system that learns from every completed job.