Instacart has acquired Arpalus, a company that uses computer vision and machine learning to translate video scans into detailed, on‑shelf inventory information, the grocery delivery giant said. The move is intended to sharpen retailers’ ability to see what is actually available on store shelves and to connect that view to online shopping and fulfillment systems.
How the technology works
Arpalus processes video from camera‑equipped devices — including smartphones and external cameras mounted on Instacart’s smart carts — and generates a near‑real‑time picture of shelf conditions. The models are designed to handle common in‑store challenges such as tightly packed products, poor lighting and intermittent connectivity.
- Accuracy: Arpalus’ scans are reported to be about 95% accurate on average.
- Scale: The technology can be used by Instacart’s network of roughly 600,000 shoppers.
- Hardware: Instacart’s smart “Caper Carts,” deployed across about 100 cities, can accept external cameras to feed live shelf data.
Why it matters
Retailers have long struggled to maintain accurate inventory visibility inside stores, which leads to missed sales, poor customer experience and inefficient fulfillment. By turning camera feeds into structured inventory signals, Instacart says it can improve in‑store execution and make ecommerce availability more reliable for consumers and retailers alike.
“We built an AI platform that gives retailers real‑time visibility into inventory and store execution, enabling faster and smarter decisions at scale,” said Ofir Zilberberg, founder and CEO of Arpalus.
Instacart framed the acquisition as a way to combine Arpalus’ vision models with its broad retail relationships and marketplace reach. The company said the shelf intelligence will be applied across its marketplace and retailer storefront products, including its Storefront Pro offering, to improve fulfillment and inventory accuracy while complying with privacy and contractual rules.
Operational impact and next steps
Arpalus’ software can prompt a shopper operating a handheld device to change camera angle or distance to capture better imagery, helping maintain data quality in real‑world conditions. The integration also allows data from cart‑mounted cameras to update on‑shelf inventory in real time, which could reduce out‑of‑stock instances and streamline online order picking.
| Metric | Reported figure |
|---|---|
| Scan accuracy | ~95% |
| Instacart shoppers | ~600,000 |
| Cities with Caper Carts | 100 |
Terms of the acquisition were not disclosed. Instacart executives described the deal as accelerating “the future of intelligent retail,” tying physical store activity more tightly to ecommerce systems and store operations.
