CES 2024: Showcasing next-generation AI accelerators
In his keynote at CES 2024, Siemens AG President and CEO Roland Busch shared his vision for an industrial metaverse that connects people, technology and AI within a highly accurate digital twin. One of the key elements to enable this is AI capable of processing and understanding complex data. Additionally, thanks to recent advances in generative AI, the way people interact with the digital world will be reshaped as well. This collaborative digital world won’t just be limited to industry either and will one day be a part of everyday life. A crucial first step on that journey is bringing AI acceleration to as many devices as possible. At CES this year, some of the worlds leading chip manufacturers, including Intel, AMD and Qualcomm, showcased the ways they’re bringing AI acceleration into wide range of devices, and what that will mean for the future.
AI in every chip
One of the major themes this year is bringing AI acceleration to the masses with all three companies featuring new purpose-built AI accelerators and NPUs in their latest generation processors aimed at consumer products, such as laptops, smartphones and tablets. As more apps and programs start to package AI functionality onboard, having a way to accelerate their inferencing on local machines without relying on the cloud or consuming a disproportionate amount of energy will be extremely important.
NPUs provide exactly this: a way to efficiently generate AI inferences at the edge in devices with constrained power profiles. While a small NPU integrated into a CPU or APU will never match the power of even a midrange GPU, much less the massive AI accelerators used in datacenters, they can still offer noticeable improvement over the low power offerings typically found in mobile products. Moving forward, having a wide ecosystem of devices fitted with AI accelerators will also be crucial in driving widespread adoption of AI-assisted technologies.
AI, especially in the consumer market, currently faces a chicken and egg problem. While some tools already require powerful hardware to run, which can then be updated to easily leverage AI functionality, bringing AI to the mainstream will necessitate it running well on average low-to-mid range hardware as well. Without the ability for it to run well on a wide range of hardware, there’s little incentive for developers to spend the time and money develop AI additions for their existing products. Likewise, without a strong software ecosystem to support it, spending valuable die space on NPUs and other accelerators is a big gamble for chip manufacturers. Now that some of the world’s largest chip manufacturers have committed to bringing AI accelerators to devices at every level of their product stacks, software adoption is sure to follow, bringing a wealth of new AI-enhanced functionality to everyday computing.
AI hits the road
Both Intel and Qualcomm have announced new chips targeted at the automotive industry. While Qualcomm is focused on bringing a unified SoC for ADAS (advanced driver-assistance system) and infotainment usage, Intel is committed to bringing AI into cars and enhancing the software defined vehicle (SDV). Generative AI offers many ways to improve natural language processing and intuitive interactions with computers. Integrating this technology into cars would allow for seamless operation of smart systems already prevalent in modern vehicles without compromising on driver awareness.
Generative AI isn’t the only type of AI which advanced chips could accelerate in cars either. Smarter ADAS systems, driver and passenger monitoring systems and, of course, self-driving and autonomous capabilities all rely on AI to some degree. As manufacturers seek to make cars not only safer, but smarter too relying on AI will be a natural step forward when it comes to creating vehicles better able to support drivers and passengers.
AI redefines reality
Bringing AI to more devices means connecting people and technology in a way that was never possible before, a bridge not just between the real and digital worlds but between people too. This revolution starts at the chip level, enabling AI in devices from the mundane to the extraordinary to deliver a truly next-generation experience. Over the next few years, AI hardware and software will advance at a rapid pace, making what today seems like fantasy common place in record time, leading us all to rethink the way we live our daily lives.
Siemens Digital Industries Software helps organizations of all sizes digitally transform using software, hardware and services from the Siemens Xcelerator business platform. Siemens’ software and the comprehensive digital twin enable companies to optimize their design, engineering and manufacturing processes to turn today’s ideas into the sustainable products of the future. From chips to entire systems, from product to process, across all industries. Siemens Digital Industries Software – Accelerating transformation.
