Problem

Almond is addressing the gap between robots that work in demonstrations and robots that can operate reliably in real factories and warehouses. The company says its earlier deployments in factories and grocery stores repeatedly broke down because data collection was difficult, reach and payload were limited, singularities constrained the workspace, and cabling and support were unreliable. This matters because much of manufacturing robotics still relies on hard-coded movement, making systems brittle when tasks or parts change.

The clearest use case is high-mix picking: identifying and manipulating different-shaped SKUs mixed together in a bin, as well as related machine-tending and inspection tasks. Conventional systems often require parts to be processed in batches and the robot to be reprogrammed between them. Almond positions its alternative as reducing the integration burden, capital expense, and deployment risk associated with traditional automation, with a pay-per-hour model intended to offer clearer ROI.

Product / Service

Almond’s core product is Axol, a dual-arm robot for physical-AI builders and real-world manipulation. It is sold directly from $7,999 and combines two 7-DOF arms, 860mm reach, 6.5kg peak payload per arm, grippers, camera interfaces, and an open-source SDK. The software stack includes bimanual inverse kinematics, VR teleoperation, data recording, and LeRobot bindings, allowing teams to build, train, and deploy manipulation policies rather than merely run fixed industrial routines.

The delivery model combines an off-the-shelf robot platform with deployment support and services. Almond offers mounting, perception and compute configurations, custom hardware integration, on-site repair, and labeled manipulation-data collection priced by usable data hours. In manufacturing deployments, its vision pipeline can combine foundational and purpose-built models, with simpler workloads running locally and more complex inference handled in hosted cloud GPUs. The intended benefit is a more adaptable robot that can handle variable, contact-rich tasks while giving robotics teams a ready-to-use physical-AI platform.

Market

Almond competes in AI robotics and physical AI hardware, spanning manufacturing automation, warehouse automation, robot-learning infrastructure, and manipulation research. Its alternatives vary by customer: ABB’s dual-arm YuMi is an established option for assembly, Universal Robots’ cobots address machine tending and material handling, and Franka’s research systems target robotics and AI developers. These are not all direct substitutes for Axol, but they represent the incumbent industrial and research platforms Almond must displace or complement; Almond’s differentiation is its dual-arm, open-source, AI-first design and relatively low entry price.

The company appears to be in early commercial deployment rather than clearly pre-revenue. Y Combinator lists Almond as an active Spring 2025 company and says Axol was available immediately, with first orders delivering that week; Almond’s own site says robots ship from San Francisco within one week. A 2026 PitchBook profile labels the company as generating revenue and reports $500,000 raised, but public materials do not disclose a customer count or revenue total. The available evidence therefore supports initial product traction and commercial availability, not yet proven scale.