An Astra-nomical Alliance

Much ink has been spilled in recent years over the massive energy consumption of many state-of-the-art artificial intelligence (AI) applications. And rightly so — when major players in the field start discussing the possibility of bringing their own nuclear power plants online to support their data centers, it is hard not to take notice. A growing number of voices have begun to ask if it is all worth it. Are the benefits of AI to humans worth the environmental impacts associated with the energy consumption of the applications?

At this point, good arguments can be made on both sides of the issue. But it is important to remember that we are not at an end state. The future of AI is much smaller, more local, and far more energy-efficient, but we have to do the dirty work of building the tools in the first place before we can turn our attention to optimization and efficiency. As we progress in these areas, the benefits of AI are expected to more clearly outweigh any environmental impacts. And that sort of work is already in full swing and producing some tangible results.

Let's get together

As a case in point, Synaptics Incorporated has just announced a collaboration with Google to advance Edge AI for the Internet of Things (IoT). Together, they aim to optimize multimodal processing for context-aware computing by integrating Google’s machine learning (ML) core, which complies with the MLIR (multi-level intermediate representation) compiler, with Synaptics' Astra hardware.

The Synaptics Astra platform is designed for IoT applications and combines scalable, energy-efficient silicon with easy-to-use, open-source software, a robust partner ecosystem, and wireless connectivity. Synaptics leverages its expertise in neural networks, AI hardware, and compiler design to support diverse modalities on the device Edge. Meanwhile, Google’s efficient ML core enhances this collaboration by meeting the demanding requirements of Edge AI devices, including power efficiency, performance, cost, and compact design.

This collaboration seems to have all the ingredients necessary to accelerate the development of AI-powered IoT devices. With their unique capabilities, these devices are able to process a variety of data modalities, including vision, voice, sound, and images, enabling intelligent interactions in applications such as wearables, home appliances, entertainment systems, embedded hubs, and monitoring tools across industries like consumer electronics, automotive, enterprise, and industrial systems.

With a shared vision to disrupt the Edge IoT space using open frameworks, these heavyweights in the world of AI may bring the next generation of powerful and efficient AI applications to us sooner than anyone had expected.