AI agentCinematic500K-1MB2BSeries C+Feature launchUS
TLVC Rating
2.40
Hook
Editing / Creativity
Copy
Sentiment of launch
Distribution strategy
Community Rating
No ratings yet
to rate this launch.
Figure AI is unveiling Index, a crowdsourced data collection platform that pays people to record video of themselves doing ordinary household and workplace tasks so those recordings can train humanoid robots. The app is now available on Google Play and the App Store , and anyone can sign up as a "Creator" to film chores like making beds, folding laundry, serving guests at a local cafe, or stocking retail shelves . The dataset feeds Helix, Figure's vision-language-action model, and is aimed at the generalization problem holding back general-purpose humanoids in the home.
The launch matters because Figure is publicly committing to a data-first approach after concluding vendor-supplied data was insufficient. Prior to this the company tried buying data, but vendors couldn't hit the throughput, diversity, or quality bar Helix requires, so Figure built the pipeline itself . Reported traction includes 264,000 downloads across 108 countries, over 44,000 weekly active users, and 30 minutes of video processed every second , with $15M paid to Creators to date . The effort began last September as an internal project called "Go-Big" before being rebranded as Index .
Figure was founded in 2022 by Brett Adcock and is headquartered in San Jose, California, where it builds the Figure 01, 02, and 03 humanoids . Adcock is also known as a co-founder of Archer Aviation and Vettery . Alongside Index, the company is signaling a $1B spend on data and compute over the next twelve months, positioning the app as the sourcing layer for that investment and inviting operators of homes, restaurants, warehouses, and offices to participate directly in generating the training corpus.
Comments (13)
Sign in to join the discussion.
N
Nadia V.Aug 26, 2026
The Index thumbnail has more kerning issues than an intern's first Figma file. Beautiful robots, tragic type stack.
R
Raj OkonkwoAug 26, 2026
$1B on data and compute in 12 months is a spicy way to say 'please stop asking about our unit economics.'
S
Sunehra P.Aug 26, 2026
264k downloads of what exactly? The tweet buried the lede so hard I had to squint. Thread structure needs a product manager.
M
mireilleAug 26, 2026
30 minutes of video uploads per second sounds impressive until you remember my group chat does that at brunch.
B
Bea CoutinhoAug 26, 2026
Launch video pacing was tight but that voiceover sounded like it was recorded inside one of the robots. Bring a pop filter to the future.
Y
Yusuke TanabeAug 26, 2026
We had a dataset like this at Meta in 2019 called 'the security cam closet.' Same vibes, worse lighting.
T
Tomas K.Aug 26, 2026
Curious how you're handling temporal alignment across heterogeneous embodiments. Is there a paper coming or are we just vibing on scale?
K
Kaveh R.Aug 26, 2026
This is the last humanoid company that will launch without an agentic data flywheel doing the annotation. Screenshot this tweet.
A
Amara O.Aug 26, 2026
Coming out of stealth by announcing you spent $15M on data is such a flex. My CFO just fainted in solidarity.
L
Linh N.Aug 26, 2026
Building something adjacent in home robotics, about 3 weeks from beta. Would love to compare notes on annotation pipelines if anyone's around.
G
Gareth MwangiAug 26, 2026
Every 'largest and most diverse dataset in the world' claim ages like unrefrigerated sushi. Show me the eval, not the upload counter.
P
Petra S.Aug 26, 2026
Sourcing engineers for two robotics teams right now, one candidate ex-locomotion at a lab that rhymes with Snoston Bynamics. DMs open.
J
João F.Aug 26, 2026
One more thing for the roadmap: a consumer-facing app so I can watch my robot fold laundry from the airport. Print money.
Work With Us
Book an intro call to explore our suite of services. We might have a waitlist.
Figure AI
The future of home help
The future of home help
Tags
TLVC Rating
Community Rating
No ratings yet
to rate this launch.
Figure AI is unveiling Index, a crowdsourced data collection platform that pays people to record video of themselves doing ordinary household and workplace tasks so those recordings can train humanoid robots. The app is now available on Google Play and the App Store , and anyone can sign up as a "Creator" to film chores like making beds, folding laundry, serving guests at a local cafe, or stocking retail shelves . The dataset feeds Helix, Figure's vision-language-action model, and is aimed at the generalization problem holding back general-purpose humanoids in the home. The launch matters because Figure is publicly committing to a data-first approach after concluding vendor-supplied data was insufficient. Prior to this the company tried buying data, but vendors couldn't hit the throughput, diversity, or quality bar Helix requires, so Figure built the pipeline itself . Reported traction includes 264,000 downloads across 108 countries, over 44,000 weekly active users, and 30 minutes of video processed every second , with $15M paid to Creators to date . The effort began last September as an internal project called "Go-Big" before being rebranded as Index . Figure was founded in 2022 by Brett Adcock and is headquartered in San Jose, California, where it builds the Figure 01, 02, and 03 humanoids . Adcock is also known as a co-founder of Archer Aviation and Vettery . Alongside Index, the company is signaling a $1B spend on data and compute over the next twelve months, positioning the app as the sourcing layer for that investment and inviting operators of homes, restaurants, warehouses, and offices to participate directly in generating the training corpus.
Comments (13)
The Index thumbnail has more kerning issues than an intern's first Figma file. Beautiful robots, tragic type stack.
$1B on data and compute in 12 months is a spicy way to say 'please stop asking about our unit economics.'
264k downloads of what exactly? The tweet buried the lede so hard I had to squint. Thread structure needs a product manager.
30 minutes of video uploads per second sounds impressive until you remember my group chat does that at brunch.
Launch video pacing was tight but that voiceover sounded like it was recorded inside one of the robots. Bring a pop filter to the future.
We had a dataset like this at Meta in 2019 called 'the security cam closet.' Same vibes, worse lighting.
Curious how you're handling temporal alignment across heterogeneous embodiments. Is there a paper coming or are we just vibing on scale?
This is the last humanoid company that will launch without an agentic data flywheel doing the annotation. Screenshot this tweet.
Coming out of stealth by announcing you spent $15M on data is such a flex. My CFO just fainted in solidarity.
Building something adjacent in home robotics, about 3 weeks from beta. Would love to compare notes on annotation pipelines if anyone's around.
Every 'largest and most diverse dataset in the world' claim ages like unrefrigerated sushi. Show me the eval, not the upload counter.
Sourcing engineers for two robotics teams right now, one candidate ex-locomotion at a lab that rhymes with Snoston Bynamics. DMs open.
One more thing for the roadmap: a consumer-facing app so I can watch my robot fold laundry from the airport. Print money.