DeepGrove
DeepGrove builds ternary-weight AI models (-1, 0, +1) designed to run frontier-quality inference on edge and low-power devices without cloud dependency.
DeepGrove
DeepGrove builds ternary-weight AI models (-1, 0, +1) designed to run frontier-quality inference on edge and low-power devices without cloud dependency.
Executive Summary
DeepGrove is a YC S25 pre-seed startup building ternary-weight AI models for edge inference — a technically legitimate and well-timed bet on a genuine structural shift in how AI gets deployed. The core market thesis holds up: edge AI software is a $2B market growing at 29%+ CAGR, and the demand for on-device inference is real and accelerating. However, the company's core technical approach (ternary weights) is directly pioneered by Microsoft Research's open-source BitNet/Bitnet.cpp project, and the team consists of two early-career founders with strong academic backgrounds but no demonstrated track record of commercializing AI infrastructure or navigating enterprise hardware partnerships. With $500K raised, no disclosed business model, no public product, and no verified customers, DeepGrove is a high-conviction bet on two young founders in a capital-intensive space dominated by well-resourced incumbents — the biggest risk is that ternary model efficiency becomes a commodity open-source standard before DeepGrove can build a defensible enterprise moat.
Run your own diligence
Upload a pitch deck or paste any company URL to get a full AI-powered due diligence report in under 2 minutes.
Get started free →Free plan available · No credit card required