In a recent study, Prowess Consulting looked into what defines an AI PC and how it can help consumers and businesses. Because the AI landscape changes week by week, we wanted to look at the most recent announcements from Microsoft, Intel, and other major market players to understand the current state of the AI PC.
As a quick refresher, “AI PC” generally refers to professional and consumer laptops and desktops with built-in hardware and software designed to process and accelerate AI workloads that run locally on the users’ devices. For hardware, that generally means a built-in graphics processing unit (GPU) or neural processing unit (NPU), but it can also refer to AI accelerators built into the central processing unit (CPU). For software, it means applications, developer frameworks, and APIs.
NPUs Are the Big New Thing
You’re likely already familiar with CPUs and GPUs, since both have been commonly available or included in PCs for years. But recently, thanks to the explosive growth of AI, NPUs are having their big moment in the consumer and business tech world. An NPU is a specialized processor designed specifically for accelerating neural network operations and AI tasks, like speech recognition for real-time transcription, background blurring in video calls, and photo or video editing processes like object detection and generative fill tasks.
Prior to the advent and recent popularity of NPUs in PCs, integrated GPUs and graphics cards dominated for processing many AI workloads. These components are still viable for this purpose, but they are generally not as power-efficient as NPUs because GPUs are designed for processing other graphics operations in addition to AI workloads.
When multiple processors (such as a CPU, a GPU, and an NPU) are combined in a single package on a PC, they can theoretically provide more efficient overall performance. Each processor handles operations for the workloads that it is best suited to take on, freeing up the other processors—the generalization workhorse CPU in particular—to take on more background tasks. This is why many vendors are offering PCs that combine the three processors in a single package.
Three-in-One
Case in point: Intel and AMD are both taking this three-is-better-than-one approach. Intel is currently featuring its recent Intel® Core™ Ultra processors, which integrate the CPU, GPU, and NPU in a single package to better support AI features such as real-time language translation, automation inferencing, and enhanced gaming environments. For more information, see www.intel.com/content/www/us/en/products/docs/processors/core-ultra/ai-pc.html.
Similarly, AI PCs powered by 2nd Generation AMD Ryzen™ AI also integrate a CPU, GPU, and NPU into a single efficient package. AI PCs from Dell Technologies, HP, Lenovo, ASUS, and Acer can be purchased with either Intel Core Ultra processors or AMD Ryzen AI chipsets inside.
Qualcomm is getting into the race, as well. Best known for creating chipsets for Android™ mobile devices and Chromebook™ devices, Qualcomm has also been making inroads into the PC marketplace by powering Windows®-on-ARM® devices. The Qualcomm® Snapdragon® X Elite processor for PCs follows the same path as Intel and AMD by offering a CPU, GPU, and NPU in a single system-on-chip (SoC), designed and marketed for AI PCs. Qualcomm chips are known for efficiency, but they can create compatibility challenges for Windows users. This is because Windows on ARM requires apps to be rewritten for native ARM64 code for the best performance. Many apps can run in an emulation layer, but they might incur a performance hit. Devices built on the Snapdragon X Elite platform are slated to ship mid-2024, so we’ll know soon if these new chips support greater compatibility for non-optimized Windows applications, while providing additional AI features and optimizations.
Other AI PC News
Microsoft is building some new AI-focused features into its upcoming Windows 11 24H2 release. The new operating system won’t be released as a standalone update until the autumn, but several new AI PC devices are expected to ship with Windows 11 24H2 pre-installed prior to that date.
Microsoft itself announced new Microsoft Surface® Pro and Surface Laptop devices in late March 2024 (with April 8 ship dates). Interestingly, it did not release consumer versions yet—these new devices are available only for business users. They feature Intel Core Ultra 5 and 7 processors with integrated NPUs for supporting AI workloads and features, including Windows Studio Effects, and they have a dedicated Microsoft Copilot button on the keyboard. These PCs will also support many other upcoming AI features in the Windows operating system (OS), including on-device live captions, AI explorer, Super Resolution, and generative AI workloads. For more information, see www.microsoft.com/en-us/d/surface-pro-10-for-business/8v73d6qwrss1.
Apple Jumps on Board
Apple switched from avoiding use of the term AI to going all in on the moniker, likely in response to AI’s rapid adoption by chipmakers, OS and PC vendors, and the general public. Earlier this year, Apple acquired Canada-based company DarwinAI. Apple is most likely using the talent and IP from that purchase to build new AI features into its next iOS® and macOS® operating systems. However, Apple notoriously keeps its plans close to the chest, so we likely won’t know what it has planned until closer to the release dates for those operating systems or Apple’s annual Worldwide Developers Conference in June.
Competition Benefits All
Although the approach and underlying silicon varies between OEMs, the goals for consumers are similar when it comes to AI workloads: run more locally, accelerate performance, and focus increasingly on AI apps and tasks that are growing in popularity and usage, like real-time translation, summarizing content, generating written or visual content, and other productivity and creativity solutions.
The different approaches and escalating competition are ultimately a win for consumers and businesses looking to solve complex challenges, accelerate time to insights, or generate new written or visual content faster.
To Learn More
If your business is looking for guidance on how to get started with AI, check out this paper from Prowess Consulting.
To learn more about what we do at Prowess Consulting, view our latest research and follow us on LinkedIn.
