The trend among PC makers toward on-device AI capabilities will enable more personalized and productive experiences for the workforce. But power-hungry apps could derail user productivity if the underlying hardware isn’t tuned to support them.
Simply put, CIOs will need to adapt their device strategies to keep up with the rapid acceleration of AI into workers’ daily routines. “We’ve seen promises of productivity-enhancing technology for decades, but they often fell short of expectations,” says Carmen True, VP of product marketing at Qualcomm Technologies, Inc. “AI is different – it’s clearly making a difference in the daily lives of everyone in the workforce. People want it integrated into their workflows to help them get their work done faster. They want to be a part of this unique moment.”
Much of the focus on building an “AI-ready” workforce justifiably focuses on proper training and providing access to AI tools and platforms. What’s missing from this guidance is the need for future-proof, AI-ready hardware that can keep up with ever-evolving demands. Specifically, this entails equipping workers with modern PCs that have the processing power and performance required to handle complex AI tasks of today and tomorrow, on top of existing demands for longer battery life and multitasking requirements.
To truly provide AI for the masses, PCs require a newer class of silicon called a neural processing unit (NPU), a specialized chip that offloads AI and machine learning tasks from the CPU and GPU to improve performance. NPUs are a core component of the emerging generation of AI PCs, which IDC predicts will comprise nearly 60% of all PC shipments worldwide by 2027.1
NPUs reduce lag time for resource-intensive AI tasks by processing workloads on the device instead of through a cloud service. Because many workloads will run with some concurrency, it’s important to split processing between the NPU, GPU, and CPU for optimal performance and power efficiency. These capabilities allow software developers to build a variety of AI capabilities into their applications. For example, Microsoft Studio Effects uses the on-board NPU to deliver a range of enhancements for video calls, such as optimized lighting, background blur, and noise cancellation.
“All those things happen organically for the user – they’re just working in the background,” says Craig Tellalian, director, field applications engineering for Qualcomm Technologies. “It leads to a better experience.”
The real promise of AI PCs comes from being able to load small language models directly on the device to generate content and insights locally. “You can create a ChatGPT-like experience, trained with your organization’s own documents and digital assets, and run it locally on a PC,” says Tellalian. “That opens up a whole set of on-device generative AI experiences for end users. Up until now, this has been limited to the public cloud, which many organizations have not enabled for fear of their intellectual property becoming compromised.”
While specialist groups like software developers and creative teams are already seeing the benefits of genAI, AI PCs will make the technology accessible to all users, at all price points. For example, Snapdragon X Series processors are available in AI PCs priced as low as $600 – all equipped with 45 TOPS (trillions of operations per second) NPUs.
“We’re delivering the same NPU performance regardless of the CPU engine,” says Tellalian. “That’s significant for democratizing AI across the workforce.”
Snapdragon branded products are products of Qualcomm Technologies, Inc. and/or its subsidiaries.
1IDC, IDC Forecasts Artificial Intelligence PCs to Account for Nearly 60% of all PC Shipments by 2027, February 7, 2024, https://my.idc.com/getdoc.jsp?containerId=prUS51851424