Intel Movidius
Generated by GPT-5-miniExpansion Funnel Raw 70 → Dedup 0 → NER 0 → Enqueued 0
| Intel Movidius | |
|---|---|
| Name | Movidius |
| Type | Subsidiary |
| Industry | Semiconductors |
| Founded | 2005 |
| Fate | Acquired by Intel Corporation in 2016 |
| Headquarters | Dublin, Ireland |
| Parent | Intel Corporation |
Intel Movidius is a brand and product line associated with low-power visual processing units and neural compute hardware originally developed by Movidius, later integrated into Intel Corporation. The technology targeted embedded vision, computer vision, and edge inference for devices such as drones, cameras, robotics, and augmented reality. The product lineage includes the Myriad family of Vision Processing Units (VPUs) and associated software stacks for deep learning and image processing.
Overview
Movidius products focused on specialized silicon for vision acceleration, enabling inference and feature extraction on devices from DJI platforms to Google-related projects. The Myriad VPU line competed and complemented offerings from NVIDIA, Qualcomm, ARM, Apple Inc., Intel Corporation, and Xilinx in applications spanning robotics, Amazon devices, and embedded systems used by Microsoft partners. Movidius technology interfaced with ecosystems including TensorFlow, Caffe, OpenCV, and frameworks supported by research efforts at Massachusetts Institute of Technology, Stanford University, and Carnegie Mellon University.
History and Development
Movidius was founded in 2005, drawing talent with prior ties to companies and institutions such as Trinity College Dublin, Cambridge University, and venture backers including Amadeus Capital Partners and Kleiner Perkins. Early milestones included shipping vision processors used in consumer electronics alongside collaborations with Sony Corporation, GoPro, Parrot SA, and research labs at Imperial College London. In the 2010s, Movidius released the Myriad 2 VPU, adopted by projects linked to Google’s Project Tango and embedded in drones by DJI. The acquisition by Intel Corporation in 2016 aimed to bolster Intel’s presence in edge AI alongside strategic initiatives associated with Intel Labs, Mobileye, and partnerships with Nokia and Lenovo. Post-acquisition, integration efforts connected Movidius technologies to OpenVINO Toolkit and collaborations with Boston Dynamics, Ford Motor Company, and Samsung research groups.
Architecture and Hardware
The Myriad VPU architecture emphasized heterogeneous compute combining vector processors, hardware accelerators, and programmable finite-state machines inspired by architectures studied at EPFL and ETH Zurich. Myriad chips incorporated multiple SHAVE (Streaming Hybrid Architecture Vector Engine) cores, hardware vision accelerators, and neural compute units supporting fixed-point and floating-point operations comparable to work from ARM Cortex designs and academic projects at University of Cambridge. Packaging and integration schemes targeted mobile and embedded markets similar to modules from Qualcomm Snapdragon partners and competitors like NVIDIA Jetson modules. Thermal and power envelopes made them suitable for unmanned aerial vehicles like those from DJI, low-power cameras by GoPro, and wearable devices showcased alongside products from Sony, HTC, and Lenovo.
Software and Development Tools
Movidius provided toolchains and SDKs enabling deployment of models from TensorFlow, Caffe, MXNet, and PyTorch via conversion and optimization tools akin to those from NVIDIA CUDA or ARM NN tools. Post-acquisition support linked functionality into Intel’s OpenVINO Toolkit and related runtime components used by developers at AWS and researchers at University of California, Berkeley. The Myriad SDK included profilers, simulators, and drivers interoperable with middleware used by robotics platforms such as ROS and vision libraries like OpenCV. Third-party integrations were visible in projects from Google research groups, collaborations with MIT Media Lab, and developer communities around GitHub and SourceForge-hosted repositories.
Applications and Use Cases
Movidius VPUs were deployed in drones, cameras, robotics, AR headsets, and smart sensors. Notable application domains included autonomous navigation used by DJI drones, object detection in devices from GoPro and Sony, and augmented reality prototypes associated with Google initiatives. Industrial and commercial deployments leveraged capabilities in retail analytics used by partners of Intel Retail Solutions and in automotive sensing explored by firms like Ford Motor Company and research teams at Waymo-adjacent labs. Academic projects at Stanford University and Carnegie Mellon University used Movidius modules for research in simultaneous localization and mapping (SLAM), while startups in startup incubators such as Y Combinator-backed teams incorporated the hardware into robotics and IoT proofs of concept.
Market Reception and Acquisition Impact
Industry reception recognized Movidius as a key player in edge vision, drawing comparisons to offerings from NVIDIA Corporation, Qualcomm Incorporated, and FPGA vendors like Xilinx Inc.. Coverage in trade media including outlets influenced by analysts at Gartner and IDC highlighted the strategic fit for Intel Corporation’s push into edge AI and computer vision alongside acquisitions such as Mobileye and investments linked to Intel Capital. The acquisition accelerated integration into Intel’s product portfolio, influencing partnerships with Samsung Electronics, Lenovo Group Ltd., and ecosystem work involving Microsoft, Amazon, and cloud providers like Google Cloud Platform. Post-acquisition, Movidius-derived IP informed developments in Intel’s edge offerings and contributed to projects showcased at conferences including CES, Computex, and Mobile World Congress.