What the data shows

1
The 17 open roles in Embodied-AI Data Annotation were spread over 13 employers; Arctic Engines held 3 of them (17.6%), and the top three together 7 (41.2%).
Figure 3
2
An hourly rate appeared on 7 of the 17 open roles in Embodied-AI Data Annotation (41.2%); their median of $30/hr (middle half $7 to $36) is 11% above the $27/hr median for all robot training work.
Figure 4
3
On RobotTraining.jobs, an open role in Embodied-AI Data Annotation got 1.51 apply clicks a week, 13.27x the robot training average, placing it the highest among 7 kinds of work with enough open time for their demand to stand alone.
Figure 6
4
0 of the 17 open roles in Embodied-AI Data Annotation (0.0%) took applicants from any country (0.3% across all robot training roles); across the 20 countries compared, people living in the United States (10), India (6) and Brazil (4) had the most to apply to, while people in Japan, Pakistan and South Africa had the fewest (0 each).
Figure 9
5
Across 17 open roles in Embodied-AI Data Annotation, the most common title words were Annotator or labeler (11), Specialist (4) and Lead, supervisor or manager (2).
Figure 10
6
Of the 17 open roles in Embodied-AI Data Annotation with a full description, none require a driving or pilot licence, none a college degree and none a background check; where years of experience are required (4 roles), the median ask is 2 years.
Figure 11
7
Over the 31 days from 1 Oct 2026 to 31 Oct 2026, 8 new roles in Embodied-AI Data Annotation were posted and 12 taken down, and removed roles had stayed up for a median of 29 days (50 for robot training work overall).
Figure 2
8
Inside Embodied-AI Data Annotation, roles listed during the window averaged 2.34 apply clicks per role a week; roles posted over 12 weeks ahead of the window averaged 0.05.
Figure 7

Figure 1

Open roles each week

Are open roles in Embodied-AI Data Annotation climbing or dropping from week to week?

Week	Open roles (fixed panel)	Index	All robot training work (index)	All tracked employers
1 to 4 OctPartial week	11	108.1	97.4	11
5 to 11 OctPartial week	10	100.0	100.0	10
12 to 18 OctPartial week	7	69.1	100.1	9
19 to 25 OctPartial week	8	79.4	101.6	12
26 to 31 OctPartial week	9	90.9	101.6	13

Base: Daily open roles in Embodied-AI Data Annotation averaged across each week, 1 Oct 2026 to 31 Oct 2026, on a fixed panel of 58 employers tracked since 1 Oct 2026 (9 of which list this work).

Index: the panel's open roles relative to the week of 5 Oct 2026, set to 100. The same index is shown for all robot training work on those employers.

Mind Robotics (first tracked 7 Oct 2026), TransPerfect DataForce (first tracked 13 Oct 2026), Applied Intuition (first tracked 14 Oct 2026), Sovrano AI (first tracked 19 Oct 2026), Alignerr (first tracked 22 Oct 2026), Latitude AI (first tracked 28 Oct 2026), HireArt (first tracked 28 Oct 2026) and Arctic Engines (first tracked 30 Oct 2026) list this work, but their tracking began after the panel was fixed. They sit outside the panel and only in the last column, so a newly tracked employer never reads as growth.

Every week is flagged: the kind moves no differently from the rest of robot training work beyond what chance would produce, so no week stands alone.

Credit: RobotTraining.jobs Research, Embodied-AI Data Annotation: Hiring and Pay Market Report, October 2026 edition.

Figure 2

Listings added and removed

How quickly do roles in Embodied-AI Data Annotation come and go, and are the same employers holding more of them?

Work	Live on 1 Oct 2026	Live on 31 Oct 2026, same employers	Change	New listings	Removed	Median days listed
Embodied-AI Data AnnotationSmall sample	11	10	-9.1%	8	12	28.5 days
All robot training work	706	714	+1.1%	175	202	50.2 days

Base: 29 roles in Embodied-AI Data Annotation appeared in listings from 1 Oct 2026 to 31 Oct 2026, out of 1,352 across all robot training work.

Change sets the window's first-day count against its last-day count, using only employers already tracked on day one, so a newly added employer cannot pass for growth.

The kind's change carries a small-sample flag: it moves no differently from the rest of robot training work beyond what chance would produce.

A new listing first appeared in the window on an employer already tracked before that week; an employer's first week (its back catalogue) is not counted.

Median days listed measures, for removed roles with a known listing date, the time from listing to the last day seen. It needs at least 3 such roles; this kind has 9.

Credit: RobotTraining.jobs Research, Embodied-AI Data Annotation: Hiring and Pay Market Report, October 2026 edition.

Figure 3

Who lists it

Who posts roles in Embodied-AI Data Annotation, and what pay does each employer put on them?

Employer	Open roles	Share of open roles	Stating an hourly rate	Median per hour
Arctic Engines	3	17.6%	0.0%	n/a
Genesis AI	2	11.8%	0.0%	n/a
Welo Data	2	11.8%	100.0%	n/a
Alignerr	1	5.9%	100.0%	n/a
Dyna Robotics	1	5.9%	100.0%	n/a
Figure AI	1	5.9%	100.0%	n/a
HireArt	1	5.9%	100.0%	n/a
Latitude AI	1	5.9%	0.0%	n/a
micro1	1	5.9%	100.0%	n/a
Mind Robotics	1	5.9%	0.0%	n/a
Nuro	1	5.9%	0.0%	n/a
Skild AI	1	5.9%	0.0%	n/a
Torc Robotics	1	5.9%	0.0%	n/a

Base: Open roles in Embodied-AI Data Annotation at the close of 31 Oct 2026 (17), split by employer. HHI 934 (unconcentrated); all robot training work: 2,197.

HHI adds up the square of each employer's percentage share of open roles, giving a score from 0 to 10,000. On the US merger-guideline thresholds, below 1,500 means unconcentrated, 1,500 to 2,500 moderately concentrated and over 2,500 highly concentrated.

Shares are of listings, not of people hired or money paid: an employer posting one role per city or country weighs more here than its hiring may.

Medians are shown from 3 roles with a posted hourly rate.

Credit: RobotTraining.jobs Research, Embodied-AI Data Annotation: Hiring and Pay Market Report, October 2026 edition.

Figure 4

What it pays

What hourly rate do roles in Embodied-AI Data Annotation post, and how does that stack up against all robot training work?

Work	Median per hour	25th percentile	75th percentile	Stating an hourly rate	Roles stating pay
Embodied-AI Data Annotation	$30	$7	$36	41.2%	7
All robot training work	$27	$23	$33	30.3%	349

Base: 17 roles in Embodied-AI Data Annotation open on 31 Oct 2026, 7 with an hourly rate, beside 1,150 open across all robot training work.

Pay is the hourly rate printed in each posting, taking the midpoint when a range is given. The 25th and 75th percentiles bound the middle half.

Per-task and per-project roles, and roles with no figure at all, count as giving no hourly rate. A median takes 3 or more priced roles.

Credit: RobotTraining.jobs Research, Embodied-AI Data Annotation: Hiring and Pay Market Report, October 2026 edition.

Figure 5

Spread of posted hourly rates

What portion of roles in Embodied-AI Data Annotation reach $50 an hour or higher?

Advertised rate	Open roles	Share of priced roles	This rate or less	All robot training work
Under $15/hr	3	42.9%	42.9%	5.2%
$15 to $25/hr	0	0.0%	42.9%	24.9%
$25 to $35/hr	2	28.6%	71.4%	50.7%
$35 to $50/hr	1	14.3%	85.7%	12.0%
$50 to $75/hr	0	0.0%	85.7%	2.0%
$75 to $100/hr	1	14.3%	100.0%	5.2%
$100 to $150/hr	0	0.0%	100.0%	0.0%
$150/hr or more	0	0.0%	100.0%	0.0%

Base: Open roles in Embodied-AI Data Annotation with an hourly rate on 31 Oct 2026: 7, set against 349 across all robot training work.

Every role is placed once, by the midpoint of its posted range, into the same bands the pay-transparency report uses.

Credit: RobotTraining.jobs Research, Embodied-AI Data Annotation: Hiring and Pay Market Report, October 2026 edition.

Figure 6

Apply clicks each opening draws

Compared with the typical robot training role, does each opening in Embodied-AI Data Annotation attract more applicants or fewer?

Roles	Weekly clicks per open role	Index	Clicks	Avg open roles
Embodied-AI Data Annotation	1.51	13.27x	73	11
All robot training work	0.11	1.00x	470	930
Its roles with a posted hourly rate	2.42	21.17x	61	6
Its roles with no posted hourly rate	0.52	4.58x	12	5
Its roles open worldwide	7.75	67.88x	44	1
Its country-limited jobs	0.68	5.98x	29	10

Base: 73 apply clicks on individual roles in Embodied-AI Data Annotation on RobotTraining.jobs from 1 Oct 2026 to 31 Oct 2026, measured against the days those roles stayed open.

Weekly clicks per role are apply clicks divided by open role-days over 7. An index of 1.00 equals the rate across all roles on the site.

Embodied-AI Data Annotation set against every kind of work: A row with under 22.2 open role-days (3.2 role-weeks) carries a small-sample mark and is left unranked: under that, the pooled rate of the kinds of work says more than the row's own clicks do (a weight fitted on those clicks).

Its roles split by posted pay, one side against the other: A row with under 9.9 open role-days (1.4 role-weeks) carries a small-sample mark and is left unranked: under that, the pooled rate of the two sides says more than the row's own clicks do (a weight fitted on those clicks). Split by reach: A row with under 2.3 open role-days (0.3 role-weeks) carries a small-sample mark and is left unranked: under that, the pooled rate of the two sides says more than the row's own clicks do (a weight fitted on those clicks). All robot training work is a single set and reads from its first open role-day.

Apply clicks reflect interest from RobotTraining.jobs visitors, who are shown roles open in their country, and recent, priced roles, first. No click figure is ever broken out by employer.

Credit: RobotTraining.jobs Research, Embodied-AI Data Annotation: Hiring and Pay Market Report, October 2026 edition.

Figure 7

Apply clicks by how old a listing is

Do fresh listings in Embodied-AI Data Annotation pull in more applicants than ones that have been up for weeks?

Listing age	Weekly clicks per open role	Index (this work = 1)	Clicks	Avg open roles
Listed during the window	2.34	1.55x	12	1
Posted 0 to 4 weeks earlier	1.51	1.00x	19	3
Posted 4 to 12 weeks earlier	3.67	2.42x	24	1
Posted over 12 weeks earlier	0.05	0.04x	1	4
Listing date unknown	3.33	2.20x	17	1

Base: Apply clicks on roles in Embodied-AI Data Annotation from 1 Oct 2026 to 31 Oct 2026, grouped by how long before or after the window opened (1 Oct 2026) each role was listed.

A role's listing date is the posting date shown on the job board, falling back to its first day in the feed. Back-catalogue roles with no date of their own have no listing date.

Here the index sets each row's rate against all roles in this kind of work, so it compares within the kind only.

A row with under 4.4 open role-days (0.6 role-weeks) carries a small-sample mark and is left unranked: under that, the pooled rate of the listing ages says more than the row's own clicks do (a weight fitted on those clicks).

Credit: RobotTraining.jobs Research, Embodied-AI Data Annotation: Hiring and Pay Market Report, October 2026 edition.

Figure 8

Country rules on who can apply

Are roles in Embodied-AI Data Annotation open to applicants in every country?

Location rule	Open roles	Share of its open roles	All robot training work
Open worldwide	0	0.0%	0.3%
Worldwide except listed countries	0	0.0%	0.0%
Only listed countries	17	100.0%	99.7%
No rule stated	0	0.0%	0.0%

Base: 17 open roles in Embodied-AI Data Annotation as 31 Oct 2026 closed, beside 1,150 across all robot training work.

Rules come from the location line of each posting. When that names no country, the posting's description is checked, and when both are silent the employer's own statement on where it hires is used.

Credit: RobotTraining.jobs Research, Embodied-AI Data Annotation: Hiring and Pay Market Report, October 2026 edition.

Figure 9

Country by country: roles a resident can apply to

For someone living in each of these countries, how many roles in Embodied-AI Data Annotation are open to them?

Country	Open roles	Share of its open roles	Median per hour
Open to every country	0	0.0%	n/a
United States	10	58.8%	$32
India	6	35.3%	n/a
Brazil	4	23.5%	$7
Mexico	3	17.6%	$4
Philippines	3	17.6%	$4
Canada	2	11.8%	n/a
Indonesia	2	11.8%	n/a
Malaysia	2	11.8%	n/a
Argentina	1	5.9%	n/a
Australia	1	5.9%	n/a
Colombia	1	5.9%	n/a
Germany	1	5.9%	n/a
Spain	1	5.9%	n/a
France	1	5.9%	n/a
United Kingdom	1	5.9%	n/a
Italy	1	5.9%	n/a
Netherlands	1	5.9%	n/a
Japan	0	0.0%	n/a
Pakistan	0	0.0%	n/a
South Africa	0	0.0%	n/a

Base: 17 open roles in Embodied-AI Data Annotation on 31 Oct 2026, checked against the same 20 countries in every report.

Each country's total adds worldwide roles, roles whose allowed list includes it, and roles whose excluded list leaves it out. Roles that state no rule at all are shown to everyone on the job board, so they count for every country here as well.

Every report in this series uses these 20 countries: the ones the location rules of open robot training roles named most often on 9 Oct 2026, whether as the only countries allowed or as exclusions.

Median pay covers the roles open to that country with a posted hourly rate, and needs at least 3 of them.

Credit: RobotTraining.jobs Research, Embodied-AI Data Annotation: Hiring and Pay Market Report, October 2026 edition.

Figure 10

What the roles are called

What are roles in Embodied-AI Data Annotation called: operator, annotator, technician or specialist?

Title word	Open roles	Share of its open roles	Median per hour	Employers
Annotator or labeler	11	64.7%	$4	8
Specialist	4	23.5%	n/a	4
Lead, supervisor or manager	2	11.8%	n/a	1
Test or evaluation operator	2	11.8%	n/a	2
Data collector or contributor	1	5.9%	n/a	1
Technician	1	5.9%	n/a	1

Base: On 31 Oct 2026, 16 of 17 open roles in Embodied-AI Data Annotation (94.1%) had one of these words in the title, with 6 title words named overall.

A title can use several of these words ('Robot Teleoperator') and counts for each. The vocabulary is that of the job-titles report.

Median pay appears from 3 roles with an hourly rate.

Credit: RobotTraining.jobs Research, Embodied-AI Data Annotation: Hiring and Pay Market Report, October 2026 edition.

Figure 11

What the descriptions ask for

Do roles in Embodied-AI Data Annotation require a degree, a licence, a minimum age or experience?

Requirement	Roles	Share of described roles	All described roles
Requires a college degree (no alternative)	0	0.0%	2.7%
Requires a driver's license	0	0.0%	49.4%
Requires a CDL, Part 107 or pilot license	0	0.0%	8.3%
States a minimum age	0	0.0%	14.8%
States a weight to lift	0	0.0%	24.0%
Requires years of work experience	4	23.5%	12.9%
States weekly hours	2	11.8%	8.6%
Requires a background check	0	0.0%	17.7%
Mentions night, weekend or rotating shifts	0	0.0%	40.8%
Says the employer trains new hires	0	0.0%	20.8%

Base: Full descriptions were available for 17 open roles in Embodied-AI Data Annotation on 31 Oct 2026 (100.0% of the kind, from 13 employers and programs); 1,150 described roles across all robot training work.

Requirements are read as requirements only (a licence named as a preference does not count); hours and shifts count wherever the description mentions them. The readers are those of the requirements report.

Many employers post one description per city, so a large employer's template weighs on the shares.

A kind's figure appears only when its described roles carry more weight than the average across kinds (2 described roles; the weight is fitted from how far each kind's shares sit from those of all described roles).

Credit: RobotTraining.jobs Research, Embodied-AI Data Annotation: Hiring and Pay Market Report, October 2026 edition.

Robotics labs · Fleet operators · Data vendors · Research teams

Buy the data behind this report

This report sums up records we gather ourselves, day after day:

1,432 robot training roles followed, history included
Every role on 100 employer job boards, read each day since 12 Sep 2026: its pay, its place, and the first and last day it was up.
536 apply clicks over 28 days
Which robot training roles people really apply to, by kind of work, pay and place. The clicks happen here, so no one else can count them.
25 sourced industry datasets
Funding, revenue, deals, fleets, deployments and prices across the robot training industry, each record linked to the publication it comes from.

On offer

The raw records
Every robot training role we track: title, employer, place, kind of work, pay as stated, and the days it was first and last open.
A regular feed
New and closed roles every week or month, as file drops or through the data endpoints.
A custom cut
Any figure here by country, city, kind of work, vehicle, employer or date, or your own hiring benchmarked against the market.

Need the people behind this data for your own robots? (you contract them yourself) or (we deliver the robot data).

Investing in robotics, AV or robot-data companies? The Robot hiring radar is these records role by role, every day, as files.

Robot hiring brief: free, once a month

The robot-training job market in figures, on the first weekday of each month: open roles, what opened and closed, who added and who cut, and posted pay by kind of work. To your work email.

How it was measured

Every day this site reads the public job boards of the robot training employers and programs it follows. Apply clicks come from visitors on RobotTraining.jobs, and full descriptions only from employers whose boards post them. A single database query produces each figure for the window shown, and the live page and every monthly edition run that identical query. The note under each figure gives its base.

Kind of work: Every role is sorted, using its title and description, into one of the seven kinds of robot training work followed here. This report covers the roles sorted into Embodied-AI Data Annotation.
Fixed panel: Only employers this site already followed when the weekly series starts. Tracking just their roles keeps a later addition from looking like growth.
Index: Average open roles across the panel in a given week, over the base week's average, multiplied by 100. It is plotted next to the same index for all robot training work at those employers.
Open role: On a given day, a role that has been posted and still shows in the employer's feed.
New listing: A role seen for the first time during the window, at an employer this site was already following the week before. Whatever an employer has up in its first tracked week is back catalogue and does not count as new.
Removed listing: A role that dropped out of its employer's feed within the window.
Days listed: For a role that came down: the span from its listing date to the final day its feed showed it. A median needs 3 or more such roles.
HHI: Herfindahl-Hirschman Index: square each employer's percentage share of open roles and add them up, giving 0 to 10,000. Below 1,500 is unconcentrated, 1,500 to 2,500 moderately concentrated, over 2,500 highly concentrated.
Advertised pay: The hourly rate a listing gives, using the midpoint when it gives a range. A median appears only when 3 or more roles state a rate, and the middle half spans the 25th to the 75th percentile.
Weekly clicks per open role: Apply clicks made on RobotTraining.jobs for a particular role, over the total days the roles stayed open, multiplied by 7.
Demand index: Weekly clicks per open role over the same rate for every role on the site. Above 1, these roles attract more applicants than a typical role.
Listing age: How long before the window's first day a role went up: inside the window, up to 4 weeks earlier, 4 to 12 weeks earlier, or longer ago. The listing date is the board's posting date, or failing that the role's first day in the feed.
Location rule: Taken from the location text of each listing or, when that text names no country, from what the listing's own description says: worldwide, worldwide with named exceptions, named countries only, or no rule given.
Open to a country: Residents of a country can apply when the role is worldwide, when it gives no rule at all (the job board treats such a role as open to everyone), when its country list includes them, or when its exclusion list leaves them out. Every report in this series compares the same 20 countries.
Window: Live page: the 28 most recent full days (UTC), plus the twelve most recent full ISO weeks for the weekly series, starting no earlier than the week of 14 Sep 2026. Edition: a single calendar month (UTC) and the weeks in it.

Limits of this data

Hiring that happens off public job boards is invisible here, and some data-collection programs sign people up inside an app instead of posting a job.
A posting does not equal a hire. Certain employers leave evergreen roles up for months, and others post the same role once per city or country.
Apply clicks show what RobotTraining.jobs visitors want, and roles placed higher get more of them. Click counts are never broken out by employer.
Requirements are shown only when this kind's described roles are numerous enough to stand on their own next to the other kinds; a weight fitted on every kind's own counts pulls a small kind toward the shares of all described roles.
When a kind of work has few open roles, a handful of listings can swing its figures, so rows resting on too little data carry a small-sample tag.

Collection and classification of listings are explained on the methodology page. Freshness, completeness and consistency are tested daily in the data quality report.

Monthly editions

The live page spans the 28 most recent full days and updates daily. Every finished month also gets its own edition, locked two days after month end, so a number quoted from it stays fixed.

Live page28 most recent full days, rebuilt daily
October 2026 edition (you are here)Will lock when the next daily refresh runs

Quoting this report

You may quote, chart or republish these figures under CC BY 4.0, provided you link back to this page. This edition is locked: its numbers will not change.

Citation

RobotTraining.jobs (2026). Embodied-AI Data Annotation: Hiring and Pay Market Report, October 2026 edition. https://robottraining.jobs/research/embodied-ai-data-annotation-market/2026-10

BibTeX

@misc{robottrainingjobs_embodied_ai_data_annotation_market_2026_10,
  author = {{RobotTraining.jobs}},
  title = {Embodied-AI Data Annotation: Hiring and Pay Market Report, October 2026 edition},
  year = {2026},
  howpublished = {\url{https://robottraining.jobs/research/embodied-ai-data-annotation-market/2026-10}},
}

Files: every figure in one CSV (a row per value) or as JSON, emailed to your work address.Other open data: datasets and API

Keep reading

Compare with every other kind of work:

Browse all 91 reports