SANTA CLARA, September 16 — Ambarella and ZEDEDA announced a partnership on Tuesday at the AI Infra Summit 2026 to bring secure, cloud-orchestrated artificial intelligence to distributed edge devices, a move the two companies say will let enterprises manage heavy AI workloads across millions of field endpoints from a single control plane. The collaboration integrates ZEDEDA’s open-source EVE-OS and Edge Intelligence Platform directly with Ambarella’s power-efficient N1 family of edge generative AI Systems-on-Chip.
In plain terms, think of it like giving every camera, robot, and connected machine in a factory its own brain — one that can be updated and supervised from a central office rather than by sending a technician to each device.
What the integration actually does
ZEDEDA’s EVE-OS, which is governed by the Linux Foundation, runs directly on Ambarella’s N1-655 silicon. That establishes a vendor-neutral operating environment capable of running containerized and Kubernetes workloads, according to the companies. Field operators can pull optimized vision models, such as Ambarella’s YOLOX, from the ZEDEDA Model Hub straight to endpoints without manual hardware intervention.
The practical upshot is that enterprises can securely deploy, update, and manage vision models, large language models, and multimodal workloads at scale across millions of distributed endpoints — all from one centralized cloud control plane. The companies say the approach is designed to reduce the cost and risk of field updates that would otherwise require sending personnel or exposing local infrastructure.
Where this lands first
The partnership targets environments that need localized, low-latency, high-reliability AI within strict power budgets. That includes robotics and autonomous machines, industrial automation and machine vision, intelligent security cameras and smart infrastructure, and automotive, mobility, and connected industrial equipment. Ambarella, based in Santa Clara, California, develops low-power system-on-chip designs for edge AI, computer vision, and video processing.
ZEDEDA, based in San Jose, California, is an edge orchestration company. The alliance combines Ambarella’s footprint of more than 50 million AI chips shipped with ZEDEDA’s orchestration platform, which the companies say is used by Fortune 500 enterprises.
Early access to EVE-OS on the N1-655 platform is open immediately. Production development kits preinstalled with the software are scheduled for release via Ambarella’s DevZone in the fourth quarter of 2026. For enterprises weighing where to run AI, the question has often been whether to send data to the cloud or keep it local.
This partnership bets on a third path — keep the computing local, but manage it as if it were in the cloud. What to watch next is whether the development kits arrive on schedule and how quickly field operators begin pulling models from the Model Hub to real endpoints.
