Y Combinator's Requests for Startups · Fall 2026

Data for the Real World

Requested by Austin Tindle & Diana Hu · View in YC · Start with a Landing Page

Build a marketing landing page for a company that collects dense, high-fidelity physical-world data — using robots, autonomous sensors, and novel methods — to train AI models for energy, agriculture, logistics, and construction. Objective: Make the case that today's physical-world data is sparse and human-designed, and that dense, machine-collected data unlocks precise modeling and control. Capture interest from design partners in the target industries. This is a landing page, not the data platform. Sensing hardware, fleets, and training pipelines happen off-platform. Core sections: 1. Hero - A clear statement: the real world is under-instrumented; we fix that - One-line thesis on dense data for real-world AI 2. The problem - Physical-world data is sparse and shaped by human assumptions - Models can't precisely predict or control what they can't densely observe - Incumbents rely on legacy, low-resolution sensing 3. The approach - Robots and autonomous sensors gathering dense, continuous data - Novel collection methods beyond fixed human-placed sensors - Turning raw physical signals into structured, model-ready datasets 4. What it enables - Precise predictive modeling and control - Concrete outcomes per industry (efficiency, yield, uptime, safety) 5. Industries - Energy, agriculture, logistics, and construction - A scenario for each showing before/after 6. Engagement - A design-partner form - Fields for industry, site type, and data needs Design: Scientific and precise — clean data-forward visuals, subtle grid and sensor motifs, and confident typography that signals rigor and scale.

Builds a marketing landing page for a company collecting dense physical-world data (via robots, autonomous sensors, and novel methods) to train AI models for energy, agriculture, logistics, and construction. Covers the thesis, the data collection approach, target industries, and a form for design partners. The sensing hardware, fleets, and model training pipeline are beyond a site builder.