Humyn Labs, a Bengaluru-based startup that records people performing everyday tasks and sells the processed footage to AI labs, raised $47 million in Series A financing, split between $26 million in equity led by Accel and $21 million in venture debt led by Lighthouse Canton Growth Debt Fund II1,4. The round is the latest in a string of large checks written specifically for companies whose sole product is data for training robots, not the robots themselves. Human Archive raised $8.2 million in July. XDOF raised $70 million in June. Humyn Labs closed its $47 million round in November. Mecka AI raised a $60 million Series B in October. MicroAGI closed a $55 million seed round in July. Five disclosed deals in five months, ranging from $8.2 million to $70 million, all channeled into a narrow category of companies that collect, process, and sell training data for embodied AI. The companies differ in method—naturalistic capture versus teleoperation logs versus synthetic augmentation—but they share a buyer base: U.S. robotics labs whose procurement budgets now carry a dedicated line item for third-party training data. That repeated, purpose-specific allocation is what makes this a funded vertical rather than a feature of the robotics stack.
Vector Wire reported on October 8 that Mecka AI's $60 million Series B, backed by Sequoia, and LG Electronics' internal data-factory effort marked the emergence of robot-training data as its own investment thesis[1]. Humyn Labs' round advances that read. The company was founded in February 2026 by Manish Agarwal, former CEO of Indian gaming company Nazara Technologies, and Ishank Gupta, a former AB InBev executive5. In roughly nine months it has reached 12 paying customers, generated what co-founder Agarwal says is $10 million in revenue between April and October, with annualized revenue he claims exceeds $50 million. Those figures are self-reported and have not been corroborated by customers or audited financials. About 90 percent of that revenue comes from U.S. buyers.
The speed matters because it shows how acute the demand signal is. Humyn Labs has developed its own two-camera hardware that records people doing their usual work or household chores rather than staged activities. It collects data across more than 20 countries in Latin America and Asia, with Brazil, India, and South Korea as its main hubs2. The company then uses its own software to process and label the footage, tagging what happens in each clip so AI models can learn from it. That pipeline—from proprietary capture hardware through annotation software to delivery—mirrors the vertically integrated approach Mecka AI described when it raised its Series B in October[1]. The structural parallel matters more than the product overlap: both companies concluded that controlling the full stack from sensor to labeled dataset is the defensible position, even though Mecka AI focuses on teleoperation logs and Humyn Labs on naturalistic video. The fact that buyers are funding both approaches suggests robotics labs are multi-sourcing across data types, not choosing one format over another.
What distinguishes Humyn Labs is the debt component. Of the $47 million, $21 million is venture debt, with Lighthouse Canton Growth Debt Fund II leading and Trifecta Capital participating. The company pays its data collectors and local partners daily or weekly but typically waits 30 to 60 days to receive payment from AI lab customers. The venture debt is earmarked to bridge that working-capital gap. Venture debt bridging a 30-to-60-day receivables gap is standard working-capital financing for any fast-growing services business with lumpy enterprise billing. What makes Humyn Labs' use of it structurally informative is the scale relative to the company's age: a nine-month-old startup already needs $21 million in debt capacity because its U.S. customers are placing orders large enough to create meaningful float. That pattern—large, recurring purchase orders from robotics labs to an outside vendor—resembles outsourced contract manufacturing more than a one-off consulting engagement, though it remains an open question whether buyers view the arrangement as permanent infrastructure or a temporary stopgap they will eventually bring in-house.
The equity side of the cap table reinforces the breadth of interest. Prosus Ventures, BeeGlobal, Bajaj Finance's Finserv Intelligence, Bajaj Alts AI Opportunities Fund, Dallas Venture Capital, Lighthouse Canton, and Jump Trading all participated alongside Accel. Former Epic CFO Anita Pramoda also joined the round. Jump Trading's presence is worth noting, though not dispositive on its own. A quantitative trading firm backing a robotics-data startup is consistent with a bet on the commodity layer underneath multiple robot platforms, but Jump has invested broadly across data and infrastructure companies, so its participation alone does not validate a commodity-layer thesis. The stronger signal is the cap table's breadth: Accel, Prosus, a Bajaj Finance vehicle, Dallas Venture Capital, and Jump represent at least four distinct investment mandates converging on the same asset class.
Humyn Labs' founders previously founded Kratos Studios, now known as KGeN. That company built a contributor network across emerging markets, and the founders said KGeN's network helps Humyn collect recordings inside people's homes. Repurposing a gig-economy network originally built for gaming into a robotics-data pipeline is an unusual origin story, but it explains how a nine-month-old company can operate across 20-plus countries without building field operations from scratch.
Humyn Labs currently has 51 full-time employees, most of them based in Bengaluru. Beyond the $10 million already booked, the company says it has another $20 million in signed contracts and a further $20 million in deals that have yet to be finalized. If those figures hold, the company's contracted and near-term pipeline would total roughly $50 million. But the $20 million in unsigned deals is the softest number in the disclosure; conversion rates on unsigned pipeline at a nine-month-old company with no public revenue track record are inherently uncertain. The $10 million in booked revenue and $20 million in signed contracts are the figures that matter for now, and even those rest on founder attestation. Accel's willingness to lead the equity at an undisclosed valuation is the strongest external signal that the revenue claims have survived diligence.
ANALYSIS The open question is whether this emerging supply base consolidates or fragments. Robotics labs today are buying from multiple suppliers because no single vendor covers every task domain and geography. Humyn Labs' 20-country footprint and proprietary hardware give it geographic breadth, but the company's reliance on venture debt to finance receivables means its growth is tethered to the payment cycles of a small number of large U.S. customers. Whether those customers begin to concentrate their spending with one or two preferred vendors, or continue to spread it across the growing field, will determine which of these startups can sustain the margins needed to service their debt. Humyn Labs' next disclosed revenue figure, and whether its $20 million in unsigned pipeline converts, will be the first concrete test.