Embodied AI startup Noematrix has raised between $30 million and $70 million in its latest funding round.

The round, labeled as pre-A++, included investors such as Shengyu Investment, Zero2IPO Capital, Vision Knight Capital, Yunqi Partners, and the Shanghai Science and Technology Innovation Group. Existing investors Prosperity7 and Sequoia China also contributed, marking their third investment in the company.

The funding will be used to enhance Noematrix’s AI models for physical environments, improve data acquisition systems, and expand commercial trials in sectors like retail fulfillment, household robotics, and food processing.

Founded in late 2023, Noematrix has completed four funding rounds. The startup was co-founded by robotics specialists Lu Cewu and Wang Shiquan, both affiliated with Stanford University.

Noematrix’s flagship product, Noematrix Brain, integrates task planning, environmental perception, and autonomous navigation. The company has introduced a “production-attached” data collection approach aimed at reducing deployment costs and improving scalability.

Its CoMiner data system, launched six months ago, has reportedly secured nearly 100 orders. The startup plans to enhance the adaptability of its embodied AI models and accelerate development using data collected through CoMiner. It is also seeking collaborations with industry and research institutions to establish a high-quality data infrastructure for broader commercialization.

🔗 Source: 36Kr

🧠 Food for thought

1️⃣ Embodied AI funding patterns reveal investor patience despite commercialization hurdles

Noematrix’s rapid succession of four funding rounds since its late 2023 founding reflects a broader pattern in embodied AI, where investors are deploying significant capital despite the well-documented commercialization challenges in robotics.

This funding approach contrasts with traditional robotics ventures, which struggled to secure sustained investment, as seen in the 2018 shutdowns of Rethink Robotics and Mayfield Robotics due to funding challenges and market fit issues 1.

The embodied AI sector now shows greater investor patience, with the global market projected to grow from $2.53 billion in 2024 to $8.76 billion by 2033, representing a 15% CAGR 2.

This shift coincides with the maturing AI ecosystem, where core capabilities in vision and planning have advanced sufficiently to make physical implementations more viable. For example, Figure AI secured $675 million and Physical Intelligence raised $400 million in similar early-stage funding 3.

The concentration of funding in fewer startups, evidenced by a 30% drop in deal count despite comparable total investment, suggests investors are backing teams with specific technical advantages rather than spreading bets across the sector 4.

2️⃣ Data acquisition emerges as the critical competitive advantage in embodied AI

Noematrix’s “production-attached” data collection method directly addresses a fundamental industry challenge—the costly and slow process of gathering physical interaction data required for training embodied AI systems.