A Californian company that was instrumental in multiple pioneering efforts to run machine learning on-device, particularly for voice recognition, Edge Impulse has agreed to be acquired by Qualcomm, which is aiming to boost its own development ecosystem for edge AI applications. Edge Impulse, founded in 2019 with the mission to streamline the creation of AI and machine learning models in embedded systems, says it will remain open to other platforms, while accelerating integration with Qualcomm devices and development kits.
The acquisition announcement was strategically released one day before the opening of Embedded World in Nuremberg, Germany, which is the largest global event of its kind for embedded developers. While the focus for this show is largely industrial systems (where “IoT” is still a “thing”), that is also largely the focus for Qualcomm, which just recently released its Dragonwing portfolio of connected devices, precisely targeting IoT and Edge AI for industrial robots, cameras, drones, telecommunication networks, etc. It’s an ambitious effort for Qualcomm, requiring a robust ecosystem of tools for development, which Edge Impulse already offers.
As developers are intensifying their efforts to enable edge AI applications, Edge Impulse’s platform for building machine learning models has been a strategic extension for many manufacturers of neural processors, AI-enabled MCUs, SoCs, and modules, particularly those running TinyML in ultra-low power designs. In that regard, Edge Impulse has been strategic for companies targeting hearables, wearables, home audio, and smart home.
Edge Impulse streamlines the creation of AI and machine learning models for edge hardware, allowing devices to make decisions and offer insight where data is gathered. Edge Impulse’s technology allows teams to rapidly develop production-ready solutions in weeks instead of years, using powerful automations. Since its foundation, the Silicon Valley company engaged with 170,000+ developers who created 450,000+ unique ML projects.
Its own team of 75 people has helped companies ship millions of AI-enabled devices that use sensors to get an actual understanding of the real world. And the company’s focus should not change with the Qualcomm acquisition – although one might expect that the priorities will be defined mostly by the largest applications that Qualcomm targets. And seeing the word “IoT” being mentioned is not a good omen.
According to the official acquisition announcement, Qualcomm Technologies intends to expand its leadership in AI capabilities to power AI-enabled products and services across “IoT”, although still subject to customary closing conditions. “This acquisition is anticipated to complement Qualcomm Technologies’ strategic approach to IoT transformation, which includes a comprehensive chipset roadmap, unified software architecture, a suite of services, developer resources, ecosystem partners, comprehensive solutions, and IoT blueprints to address diverse industry needs and challenges,” the release states.
“We are thrilled about the opportunity to significantly enhance our IoT offerings with Edge Impulse’s advanced AI-powered end-to-end platform that will complement our strategic approach to IoT transformation,” says Nakul Duggal, group general manager, automotive, industrial and embedded IoT, and cloud computing, Qualcomm Technologies. “We anticipate that this acquisition will strengthen our leadership in AI and developer enablement, enhancing our ability to provide comprehensive technology for critical sectors such as retail, security, energy and utilities, supply chain management, and asset management.”
Maybe it’s too early, but I predict that for companies and developers working in the consumer space, connected development will have to rely on other partners who will prioritize the very specific requirements of personal and home applications. This is also a great opportunity for many engineering service companies to expand their development platform offerings away from industrial applications, and obviously a big incentive for many companies— such as Ceva, Syntiant, Synaptics, and many sensor manufacturers— to diversify their development ecosystems and maybe also add their own development kits.
It’s true that Edge Impulse’s development platform already includes a comprehensive set of tools and features for data collection and preparation, model training, deployment, and monitoring — all with an easy-to-use, user-friendly interface requiring little or no code at all. That includes using Edge Impulse’s platform to build solutions for audio events and speech recognition applications relying on MCUs, CPUs, GPUs, and NPUs. Let’s see what happens and if Qualcomm allows Edge Impulse to grow its own team to support new efforts and sustain existing ones.
www.edgeimpulse.com
www.qualcomm.com
Source link
