Ambarella and ZEDEDA Partner to Bring Cloud-Orchestrated AI to Billions of Devices at the Physical Edge

ZEDEDA Edge Intelligence Platform now runs on Ambarella's edge AI SoCs, letting enterprises deploy, secure and operate AI models on power-efficient silicon from a single control plane

Joint demonstrations with Roboflow and Liquid AI debut at the AI Infrastructure Summit 2026

SAN JOSE, Calif. and SANTA CLARA, Calif., Sept. 15, 2026 (GLOBE NEWSWIRE) -- Ambarella, Inc. (NASDAQ: AMBA), a leading developer of edge AI semiconductor solutions, and ZEDEDA, the leader in edge intelligence, today announced a strategic partnership to bring secure, cloud-orchestrated AI to the devices where the physical world meets AI: cameras, robots, vehicles, industrial systems and edge appliances.

AI silicon has advanced faster than the tools to manage it in the field. Enterprises can build powerful edge AI hardware, but deploying, updating and securing models across fleets of devices remains largely manual, slow and hard to secure. ZEDEDA's 2026 survey of 600 IT and business leaders found that 47% of enterprises have already adopted hybrid cloud-edge architectures, and that 41% describe managing AI workloads across those environments as a challenge.

Through this collaboration, ZEDEDA's open-source EVE-OS, governed by the Linux Foundation, now runs on Ambarella's N1 family of edge generative AI SoCs, and ZEDEDA's Edge Intelligence Platform can deploy, update and operate AI models directly on Ambarella's power-efficient AI acceleration hardware. Enterprises and OEMs can now orchestrate vision models, LLMs and multimodal AI workloads on Ambarella-based systems at any scale, from a single control plane, with zero-touch security built in. This represents a significant evolution in edge computing: cloud-native software infrastructure and Physical AI inference together inside the machine, with centralized control and management retained.

The partnership unites two complementary leaders: Ambarella's SoCs power AI perception across security cameras, robotics, automotive and industrial systems, with more than 50 million AI chips shipped cumulatively. ZEDEDA's platform is trusted by leading Fortune 500 enterprises operating tens of thousands of edge nodes in the field. Together, the companies remove the biggest barrier to physical AI adoption: getting models onto distributed devices, keeping them current and keeping them secure, without sending trucks or exposing infrastructure.

Building on a breakout year for both companies

Today's announcement extends the momentum both companies have built in 2026:

• In March, ZEDEDA unveiled its Edge Intelligence Platform at NVIDIA GTC, the industry's first platform to create, secure and operate edge and physical AI at scale, alongside Edge Intelligence Labs and Edge Intelligence Appliances.
• In January, Ambarella launched its CV7 edge AI vision SoC at CES 2026, delivering more than 2.5x the AI performance of the prior generation on 4nm process technology, and opened its Developer Zone to broaden the Ambarella edge AI ecosystem.

This partnership connects those threads: Ambarella's newest silicon and developer ecosystem, now operable at fleet scale through ZEDEDA Edge Intelligence Platform and edge AI ecosystem.

What the partnership delivers

• EVE-OS validated on Ambarella N1: ZEDEDA's open-source, Linux Foundation LF Edge-based edge operating system runs on the Ambarella N1-655, providing a secure, vendor-neutral foundation for containerized, virtualized and Kubernetes workloads on-device.
• Model deployment from the ZEDEDA Model Hub: Ambarella-optimized AI models are deployed to Ambarella-based systems directly from ZEDEDA's Edge Intelligence Platform, with verified inference on Ambarella's AI acceleration engine and full observability, monitoring and lifecycle management for ongoing updates and optimization in the field.
• A full-stack developer path: Ambarella's edge AI platform combines proprietary AI acceleration and computer vision with its Cooper development environment: 12 edge AI SoC families supporting more than 200 AI model architectures in customer production and now orchestrated at fleet scale by ZEDEDA.
• Developer kits, ready out of the box: Development kits featuring Ambarella technology with EVE-OS preinstalled are expected to become available through Ambarella's DevZone, so developers and system integrators can go from unboxing to orchestrated Physical AI deployment in minutes.
• An open ecosystem path: The companies are working with ISVs, system integrators and hardware partners to deliver full-stack edge AI solutions, and will explore deeper participation in the LF Edge ecosystem, of which ZEDEDA is a founding member.
  

A phased path to enterprise scale

IDC forecasts that worldwide edge computing spending will approach $360 billion by 2027, and a growing share is going toward systems that run AI inference on the device and on edge AI boxes. To address this market, the companies have defined a five-phase strategic roadmap, and—contingent on customer adoption and on the companies electing to proceed—the roadmap includes deployments reaching hundreds of thousands of managed edge nodes across more than one hundred enterprises, with associated Ambarella revenue that could exceed hundreds of millions of dollars over a 5-7 year period. These figures represent agreed-upon planning objectives rather than minimum revenue commitments, purchase obligations or guaranteed forecasts.

The collaboration initially focuses on use cases where AI must operate locally with low latency, high reliability and constrained power budgets, including robotics and autonomous machines, industrial automation and machine vision, intelligent cameras and smart infrastructure, automotive and mobility, retail and logistics, and connected industrial equipment.

See it live at AI Infra Summit 2026

The companies will showcase the integration at the AI Infra Summit, September 15-17 at the Santa Clara Convention Center, where Ambarella is a Diamond Sponsor. Running on the Ambarella N1-655 platform, the demonstration will show how enterprises can create, package and deploy a custom solution blueprint to an Ambarella cluster, with the Ambarella YOLOX AI model automatically pulled from the ZEDEDA Model Hub and made available to the runtime at startup. Ecosystem partners Roboflow and Liquid AI will showcase computer vision and language models running on Ambarella silicon in the same demonstration, orchestrated by ZEDEDA.

The companies will also demonstrate Kubernetes-based AI workloads running on Ambarella CV7, illustrating how sophisticated cloud-native workloads can be orchestrated directly inside highly power-efficient intelligent edge devices — including camera-class systems. This represents a significant evolution in edge computing: bringing cloud-native software infrastructure and Physical AI inference together inside the machine while retaining centralized control and management.

Executive quotes

"AI is moving into the physical world, and the hard part was never training models. It's operating them securely and reliably on millions of devices in the field," said Said Ouissal, founder and CEO of ZEDEDA. "Ambarella has built an exceptional AI compute platform for the physical world. By combining Ambarella silicon with EVE-OS and ZEDEDA Edge Intelligence Platform, we can give developers and enterprises a cloud-like experience for deploying and operating AI across fleets of intelligent machines. This is an important step toward making Physical AI operational at scale."

“Physical AI is moving intelligence from the cloud into machines that perceive, understand and act in the real world,” said Muneyb Minhazuddin, Chief Growth Officer at Ambarella. “Ambarella provides the AI compute platform that makes this possible within extremely power-constrained devices, while ZEDEDA provides the cloud-like infrastructure to securely deploy, orchestrate, observe and manage that intelligence at scale. Together, we are creating a powerful foundation for the next generation of intelligent cameras, robots, vehicles and industrial machines.”

Availability

EVE-OS support for the Ambarella N1-655 is available now for early access customers. Validated Ambarella solution blueprints and optimized models in the ZEDEDA Model Hub, and development kits with EVE-OS preinstalled, are expected to be available in Q4 2026. The companies are currently extending the integration with Ambarella hardware observability and direct Model Hub deployment, creating a path toward repeatable, fleet-scale deployment of Physical AI applications. Contact ZEDEDA or Ambarella for early access.

About ZEDEDA

ZEDEDA unlocks the value of AI where it matters most, enabling enterprises to create, secure and operate edge AI at scale. ZEDEDA's Edge Intelligence products and solutions are used by globally distributed enterprises to rapidly realize intelligence, with real-time data driving business outcomes. Trusted by the world's largest organizations, ZEDEDA is backed by world-class investors and has teams in the United States, Germany, India, and the United Arab Emirates. For more information, visit ZEDEDA.ai.

About Ambarella

With an installed base of more than 50 million AI SoC units, Ambarella’s products are utilized in a wide variety of physical edge AI applications, spanning edge endpoint and edge infrastructure use cases including physical security, vehicle safety, telematics, autonomy, portable video, aerial drones, and other emerging robotic applications. Building on this footprint, Ambarella offers a full-stack edge AI platform, from highly optimized silicon and programmable software to AI agentic frameworks that coordinate perception, decision-making and control across devices. Ambarella’s low-power systems-on-chip (SoCs) integrate proprietary and highly efficient perception and deep learning neural network AI accelerators, enabling electronic systems to become more productive with partial or complete levels of machine autonomy. For more information, please visit www.ambarella.com.

Media Contacts

ZEDEDA:
Treble
Sarah Vandiver
zededa@treblepr.com

Ambarella:
Jonathan Miller
jmiller@ambarella.com