Problem

Sureform addresses the real-world data bottleneck in multimodal and embodied AI. Robotics teams building world models and robot foundation models need massive quantities of task-specific human and robot data, but the relevant examples are specialized, physical, and difficult to capture. The killer use cases are tasks such as restocking inventory in a pharmacy or repairing downed power lines—activities where robots must learn from real human expertise rather than generic internet data.

The pain is both scarcity and economics: conventional teleoperation-based collection can require a physical robot costing $40,000 or more per operational platform, while lab- or office-based collection produces less varied and realistic demonstrations. Sureform’s opportunity is to make specialized workplace activity a scalable source of training data by capturing what employees already do in hotels, warehouses, restaurants, construction sites, and similar environments.

Product / Service

Sureform operates as a managed, B2B data-collection platform connecting workplaces with robotics and AI teams. Employees are equipped with head-mounted RGB cameras, stereo-depth rigs, and handheld grippers while they perform real tasks; the system captures time-synchronized video, audio, pose, depth, and tactile data. Each session is structured, synchronized, and delivered training-ready, so robotics teams can obtain task-specific multimodal datasets without building the entire collection operation themselves.

The benefit is richer and more realistic training data for models that must understand and act in the physical world. For participating businesses, Sureform also presents data collection as a potential additional revenue stream and a path toward eventually automating physical operations with robots; for data buyers, it offers access to workplace-derived human demonstrations tailored to particular model or task requirements.

Market

Sureform competes in the emerging market for multimodal human data, embodied-AI data infrastructure, robotics training data, and AI data marketplaces. Its closest apparent peers include Sensei, which outsources robotic training-data collection through a low-cost hardware platform and operator network; Cortex AI, whose marketplace pays workplaces to host data-collection sessions; and Nferent AI, which offers real-world multimodal capture, teleoperation, annotation, and standardized datasets for robotics labs. Sureform’s distinctive angle is the direct use of ordinary workplaces and employees to generate task-specific, synchronized data in natural operating environments.

The company is early-stage rather than a demonstrated scaled vendor. Y Combinator lists Sureform as founded in 2025, part of its Spring 2025 batch, active, San Francisco-based, and a two-person team. Public materials reviewed do not disclose revenue, named customers, pricing, or collection volume; separate startup directories report $0 raised or an unfunded stage. The best-supported characterization is therefore pre-scale and likely pre-revenue, with YC participation and an operating workplace/data-collection concept as its primary publicly visible traction rather than disclosed commercial metrics.