Edgify Raises $9m to Scale Edge AI Infrastructure for Physical Retail

Edgify, a London-based edge machine learning operations platform for physical retail, has raised $9 million (approximately €7.7 million) in Series A+ funding. The round was backed by Rank Ventures and Mangrove Capital Partners and brings the company's total funding to $25 million (approximately €21.6 million).

Founded in 2019, Edgify connects and trains machine learning models across in-store devices including self-checkouts, cameras, scales and point-of-sale terminals. The company says the approach allows hardware to learn locally and share what it learns across a network without raw data leaving the store, reducing cloud infrastructure cost and latency while keeping customer and operational data on site.

What the funding will support

Edgify said the capital will fund the rollout of its platform and expansion beyond grocery into quick-service restaurants, distribution centres and apparel retail, and over the coming months it intends to extend the platform to manage the full lifecycle of AI models across physical retail estates.

Loss prevention is the company's initial commercial application: the platform recognises produce and identifies scan avoidance, product substitution and cart-based loss at self-checkout. Edgify states that its platform is hardware-agnostic and runs across existing in-store equipment from multiple manufacturers, including Zebra Technologies and Bizerba. It works with grocery retailers and technology partners in the United States and Europe.

Chief executive and co-founder Nadav Israel said the company was built on the conviction that intelligence should sit where data is created, and that the operational problems solved in grocery and convenience retail apply in any industry running fleets of devices in physical environments. Chief operating officer Mitchell Goldman said retail is a demanding test case for edge AI because of device density and real-world constraints on cost, latency and privacy.

Why it matters for private capital

Enterprise AI investment has concentrated on cloud-hosted models, but a growing share of deployment is constrained by connectivity, latency and data residency rules. Infrastructure that runs inference and training on existing hardware addresses those constraints without new capital expenditure by the customer — a proposition that has drawn investor attention across retail, logistics and manufacturing. Adoption in this segment is typically slow and estate-wide, which lengthens sales cycles but tends to produce durable contracts once installed.