Human Archive

Multimodal data provider for robotics learning

humanarchive.ai

HQ: San Francisco, CA, USA

1-10

Industrials

Company research

Public company, workplace, funding, and market signals

Updated Aug 16, 2026

Overview

Human Archive is a YC W26 robotics/physical-AI data company building multimodal, human-embodied intelligence datasets for robotics and world models. Its core product appears to be a data platform plus hardware-enabled data collection infrastructure that captures synchronized egocentric video, depth, tactile, motion, and audio signals from real-world human work environments.

Primary product

Multimodal physical-AI training data platform and hardware-enabled data collection system for robotics/world models

Founded

2026

Headquarters

San Francisco, California, United States

Team size

10-20 employees

Work style

Flexible

Industry

Robotics Engineering

Sub-industry

Physical AI / robotics training data / embodied intelligence datasets

Offices

San Francisco, California, United States

Bengaluru, India

China (hardware lab/opened in Shenzhen, per LinkedIn post)

Southeast Asia (expansion mentioned in reporting)

Business model

B2B data licensing

data collection services

hardware-enabled data platform

potential dataset/API-style access for frontier AI labs

Funding

Stage

Seed

Total raised

$8.3M

Latest round

Seed · May 2026

Latest amount

$8.2M

Seed

May 2026 · Wing Venture Capital

$8.2M

People & investors

Leadership

Raj Patel

CEO / Co-Founder

Rushil Agarwal

Co-Founder

Samay Maini

Co-Founder & CTO

Shloke Patel

Co-Founder

Mohd Tanveer

Founding Ops Team

Mihir Parikh

Ops Partnerships

Investors

Work & culture

Public posts suggest a very early, high-agency, intense startup culture: lean team, rapid hiring, founder-like thinking, strong execution bias, and willingness to operate without much structure. Public statements also emphasize research depth and speed.

Work style

Flexible

Compensation

No formal public compensation policy found. Public hiring posts suggest aggressive early-startup pay for operations roles, including a post mentioning up to ₹1,00,00,000/year for founding ops in India.

Product & market

Pricing

Not publicly disclosed; likely enterprise licensing / custom data contracts / services-based engagement.

Differentiators

—Synchronized multimodal capture beyond video alone
—Physical-world, in-situ data collection rather than lab-only or synthetic-only data
—Proprietary hardware rigs deployed across real work environments
—Focus on egocentric human sensorimotor data for embodied AI
—Early scale claims across many environments and geographies

Technology

wearable cameras

RGB-D / depth sensors

tactile sensors / force feedback

motion capture

wrist cameras

chest cameras

anonymization pipelines

annotation pipelines

benchmarking pipelines

internal model training

robot policy training

Customers

Frontier AI labs

Robotics companies

World-model researchers

Potentially U.S. household/service and labor-network pilots

Competitors

Scale AI (physical AI/data engine)

Build AI

other robotics/embodied-AI data vendors

in-house data collection teams at robotics labs

Scale & contact

Estimated revenue

Unknown; based on seed-stage status, likely pre-scale or low seven-figure ARR at most, but no reliable public evidence found.

Signals & risks

Other

May 2026

Public scrutiny emerged around worker consent, pay, and data collection practices after media coverage of workers wearing head-mounted cameras in India.

Source

Other

May 2026

Reporting noted public debate and regulatory attention in India around the company’s gig-worker data collection model and consent practices.

Source

Latest news

Human Archive Raises $8.2M in Seed Funding

FinSMEs · Jun 2026

Confirms the seed round and frames Human Archive as a data platform for human embodied intelligence, with use of proceeds for engineering, hardware, and operations.

Human Archive Raises $8.2M for Physical AI Training Data

The Robotics Media · May 2026

Summarizes the seed raise and the company’s multi-sensor training-data infrastructure.

Human Archive Raises $8.2M in Seed Round Funding to Model Human Embodied Intelligence to Train Robots

The AI Insider · May 2026

Reported the funding round, the multimodal wearable-data approach, and the company’s expansion across hardware, software, and operations.

Human Archive Wants Data to Train Physical AI—By Whatever Means Necessary

Analytics India Magazine · May 2026

Covers the privacy controversy, the physical-AI market thesis, and reported worker-compensation ranges in the broader debate around consent and surveillance.