Humyn Labs, a physical AI research and deployment company, announced USD 47 million in Series A funding to accelerate robot deployment-readiness across real-world environments in Brazil, South Korea, and India. Humyn Labs says it turns human experience into physical understanding for robots by capturing real-world work across factories, warehouses, farms, and homes. The company works with frontier AI labs, robotics companies, and world-model developers across global markets.
Why Physical AI Requires Real-World Sensory Data and Testing
Gartner defines physical AI as systems that consume, understand, and interact with physical environments, bringing technical challenges not seen in digital-only AI when autonomously and safely engaging robotics in the real world. According to Arthur D. Little’s physical AI viewpoint, classical machine learning and generative AI train primarily on publicly available internet data to produce digital outputs, whereas physical AI captures real-world data through sensors and acts directly on the physical world through actuators.
According to Arthur D. Little, physical AI differs from conventional rule-based robotics because its processing engine communicates bidirectionally with a multimodal AI system, allowing devices to adapt to changing inputs and improve through learning without reprogramming. Arthur D. Little also notes that physical AI devices must be error-free before launch because failures affect the physical world, meaning testing and validation in real-world scenarios extend development timelines compared with software applications.
Series A Equity and Venture Debt Investors
Humyn Labs’ Series A announcement splits the USD 47 million financing into a USD 26 million equity round led by Accel and a USD 21 million venture debt facility led by Lighthouse Canton. Alongside Accel, the equity round drew participation from Prosus Ventures, BeeGlobal, Finserv Intelligence through Bajaj Finance, Bajaj Alts AI Opportunities Fund, Dallas Venture Capital, Lighthouse Canton, Jump Trading LLC, and angel investor Anita Pramoda, former CFO of Epic.
Trifecta Capital joined Lighthouse Canton in the USD 21 million venture debt round. According to TechCrunch’s report on Humyn Labs, Humyn Labs raised the venture debt portion for working capital because it pays workers and data collection partners daily or weekly while waiting 30 to 60 days for customer payments. Pratik Agarwal, Partner, Accel, said, “Physical AI is a defining technology shift of the next decade, and data and deployment are where the real moat will be built. Humyn Labs has the technical depth and customer pull to be a category leader.”
Capture Hardware and Multi-Sensory Data Engine
Humyn Labs runs capture and enrichment operations across 20+ countries in Latin America, East Asia, and South and Southeast Asia. Its proprietary capture devices record specific jobs to lab requirements, while its data engine converts recordings into labelled, training-ready datasets with verified provenance across sight, sound, motion, and touch. According to TechCrunch, Humyn Labs developed its own two-camera hardware to record people performing everyday tasks and uses software to tag what happens in each clip and track hand movements and object interactions.
Egocentric and Residential Datasets on Hugging Face
On Humyn Labs’ Hugging Face profile, the company publishes 12 datasets, including Egocentric-3-Camera-Array, APAC-Egocentric-Residential-Voiceover, APAC-Egocentric-Stereo-Labeled, APAC-Egocentric-Monocular-Labeled, and LATAM-Egocentric-Residential-IMU. According to TechCrunch, co-founders Manish Agarwal and Ishank Gupta previously founded Kratos Studios, now known as KGeN, whose network of contributors across emerging markets helps Humyn Labs collect data in people’s homes.
Voice AI Benchmarking with BRIDGE
In May 2026, Humyn Labs published BRIDGE (Benchmark of Regional & International Data for Global Evaluation), evaluating 15 commercial AI speech-recognition tools across 22 non-English languages spoken by over 5.5 billion people. According to Humyn Labs’ BRIDGE launch release, the company built the benchmark on real two-person conversations collected across multiple regions rather than scripted or internet-scraped audio, and made the full dataset available on Hugging Face.
Deployment-Readiness Infrastructure and Robot Gyms
The Series A funding will support deployment-readiness infrastructure in Brazil, South Korea, and India, alongside capture hardware, enrichment engines, and on-the-ground presence in San Francisco, London, and Seoul. According to Business Standard, Humyn Labs plans to establish dedicated robot deployment centres called “robot gyms” where models and hardware can be tested in real-world settings, and will work with large enterprises to build centres at their facilities.
Global Commercial Expansion Hubs
According to Business Standard, Humyn Labs is headquartered in Bengaluru and operates a U.S. subsidiary. Business Standard reports that the company is prioritising San Francisco as its first international market, followed by London, Munich, Seoul, and Tokyo.
Humyn Labs Founding, Revenue, and Investors
Humyn Labs was founded in 2026 by Manish Agarwal, an alumnus of IIM Ahmedabad, and Ishank Gupta, an alumnus of The Wharton School. According to TechCrunch, Manish Agarwal is the former CEO of Indian gaming company Nazara Technologies, and Ishank Gupta is a former executive at brewing company AB InBev. In April 2026, Business Standard reported that Humyn Labs announced a $20 million commitment to develop human data infrastructure for physical AI firms worldwide.
Humyn Labs says it has surpassed USD 50 million in annualised revenue within six months of operations. According to TechCrunch, the startup generated $10 million in revenue between April and October 2026, holds another $20 million in signed contracts, and has a further $20 million in deals not yet finalized. Manish Agarwal, Co-founder, Humyn Labs, said, “Physical AI will only be as good as the real-world data it learns from. This funding will help us strengthen our research, build trust with labs and enterprises, and expand globally. With a USD 10 billion opportunity in data infrastructure, we aim to build a USD 1 billion annual revenue business.”
Questions about Humyn Labs Series A
How does Humyn Labs collect real-world training data without staged recordings?
According to TechCrunch, Humyn Labs collects data from people doing their usual work or household chores instead of staged activities, working with local partners to identify collection sites and supervise workers. On Hugging Face, Humyn Labs describes its offering as the multi-sensory platform for physical AI covering sound, sight, mobility, and touch in one fused training signal from the real world.
Who are Humyn Labs’ customers and where is its team based?
According to TechCrunch, Humyn Labs has 12 paying customers, including frontier AI labs, robotics companies, and developers of world models, with about 90% of its revenue coming from the U.S. TechCrunch reports that Humyn Labs employs 51 full-time workers, most based in Bengaluru, alongside local partners across its collection countries.
How does the BRIDGE benchmark evaluate non-English voice AI models?
Humyn Labs states in its May 2026 release that BRIDGE evaluates speech models using seven metrics beyond standard word and character error rates. Those metrics include Semantic Similarity, Code-Switch F1, Loan Word WER, Phoneme-Informed Error Rate, and Word Information Lost.
Why is India not an immediate commercial robot deployment market for Humyn Labs?
According to Business Standard, co-founder Manish Agarwal stated that India may take longer to become a priority market for commercial robot deployment because robot cost economics do not yet work across several local use cases. While India serves as a core hub for data collection and deployment-readiness infrastructure, commercial customer expansion focuses first on U.S. and European markets.


