There is already generative AI for users through a variety of cloud-based services, from OpenAI’s ChatGPT to Google’s Gemini. Nevertheless, Forrester Research predicted that next year will be the ‘year of the AI PC.’ Forrester defines an AI PC as a PC with built-in AI processors and algorithms specifically designed to improve the AI workload experience across CPU, GPU, and NPU. NPU offloads certain functions, allowing PCs to run AI algorithms at lightning speed.
“Employees have been using AI for things like blurring backgrounds and removing noise in client operating systems for years, but most AI processing still happens within cloud services like Microsoft Teams,” Forrester said in the report. “AI PC breaks down the cloud-centric AI model and brings AI to local devices,” he explained.
AMD, Dell, HP, Intel, AMD, Apple, NVIDIA, and Lenovo are all predicting AI PC innovation over the next year or so. And this period is an important period for Windows users. With Windows 10 ending support next October, there’s now a real reason to upgrade to Windows 11 and buy new hardware.
According to Gartner’s global AI PC shipment forecast, a total of 114 million units are expected to be shipped in 2025, a 165.5% increase from this year. Gartner’s key forecasts are as follows:
- AI PCs, which accounted for only 17% in 2024, will account for 43% of total PC shipments in 2025.
- The demand for AI laptops is expected to grow higher than that of AI desktops, and shipments of AI laptops are expected to account for 51% of all laptops by 2025.
- By 2026, AI laptops, which accounted for less than 5% in 2023, will be the only laptops large companies can choose from.
“The debate is moving from speculating about which PCs will include AI capabilities to predicting that most PCs will eventually have AI NPU capabilities integrated,” said Ranjit Atwal, senior director analyst at Gartner. “As a result, NPU will become a standard feature for PC vendors.”
Atwal said that as the PC market shifts toward AI PCs, in the long term, in the consumer AI laptop market, ARM-based AI devices will take share from Windows x86 AI and non-AI laptops, reducing the dominance of x86 processors over time. expected. “But by 2025, Windows x86-based AI laptops will dominate the business sector,” Atwal added.
Why bother embedding AI algorithms in your computer’s firmware or software and need a more expensive processor to run them when you can access the same tools on the web? According to Tom Butler, general manager of worldwide commercial product management at Lenovo, AI will fundamentally transform PCs, making them not only more intelligent, but also more responsive and secure.
“AI PCs are evolving to deliver more personalized and adaptive experiences tailored to the needs of each user,” Butler said. “The emergence of generative AI was an important moment, but reliance on cloud processing raises significant concerns about data privacy,” he said.
According to Butler, each component of a PC plays a unique role in making AI tasks efficient, but NPUs play a key role in accelerating AI computations with minimal power consumption. In general, Butler explains, AI PCs assist or process users to perform everyday tasks more efficiently and intuitively without the need to access external websites or services.
For example, Apple last month announced an updated iMac powered by a new M4 chip with NPU cores and Apple Intelligence, an AI-based assistant that can help users compose emails and other content. Complex tasks can still be transferred to OpenAI’s ChatGPT. Apple also unveiled the M4 Pro-based MacBook Pro laptop and Mac mini, touting their strengths in handling AI-related tasks.
According to Butler, AI PCs can improve productivity by handling routine tasks like scheduling and organizing email, and enhancing collaboration with real-time translation and transcription features.
Depending on the device, AI technology can seamlessly integrate with cloud and edge computing for real-time data processing to enable faster, more informed decisions. AI-enabled PCs also enhance device security by automating threat detection and adapting to new threats as they arise. For example, Lenovo’s Smart Connect increases device synergy so users can seamlessly switch between Lenovo devices, while ThinkShield provides ecosystem-wide security to protect users in real time.
However, AI PCs typically require more RAM to handle advanced tasks. For example, Apple increased the minimum memory from 8GB of RAM to 16GB.
Lenovo took a slightly different approach with RAM support for AI workloads. Lenovo’s so-called ‘smarter AI for everyone’ seeks to match task complexity and processing requirements. For example, 16GB is suitable for light AI tasks when combined with a more powerful NPU, while 32GB or more is suitable for users handling complex applications, LLM, or deep learning.
“Users working in AI development will need more RAM along with powerful GPUs and CPUs to ensure low latency and the ability to fine-tune AI models,” Butler said.
Guiente San Martin, president of HP’s Commercial Systems and Display Solutions division, said PCs with NPUs capable of more than 40 TOPS per second are revolutionizing everyday workflows, with new use cases emerging almost every day. LLM tools like Microsoft Co-Pilot and ChatGPT can quickly summarize documents, draft emails, schedule meetings, and even create action items from meeting summaries to increase productivity and free up time on the job. .
“We are also seeing use cases in creative, editing, design, and content creation fields,” San Martin said. “Now, with AI, we can create images, videos, and music with just text prompts, so anyone can become a composer or designer. Now we have a product. Rather than having to wait for a creative agency, marketers can create campaign copy and fully realized campaign collateral in hours rather than weeks, even by professional video and design professionals, automating complex processes like background removal or generative extensions with a single click. You can use tools that “There is,” he emphasized.
Additionally, the emergence of AI-enhanced devices that can perform real-time language translation through live captioning, speech-to-text, and voice commands in real-time meetings can help break down barriers between global teams. AI tools like Virtual Sapiens provide real-time coaching during meetings to help users communicate more professionally and clearly.
The main advantage of running AI on the device instead of a cloud service is that security is automatically enhanced. “AI also automates control of the PC itself,” Sanmartin said. “It checks security settings, configures the camera and microphone for calls, and adjusts the user interface according to the user’s preferences or automatically based on usage patterns. “You can ask AI to adapt to improve security and threat defense by monitoring unusual patterns in your data and behavior and blocking threats in real time.”
AI PCs that support local inference (running language and visual models directly on the device) have several advantages. Personal, proprietary or financial data remains local rather than uploaded to cloud-based AI, allowing users to safely handle sensitive information. San Martin emphasized, “AI on the PC can be used for generative tasks without exposing the data and content in all personal files to external servers.”
AI makes things more difficult
But ironically, a new survey and study conducted by Intel shows that current AI PC users spend more time doing work than PC users without AI technology. A survey of 6,000 users in Germany, the UK and France found that around 53% believe AI PCs are only useful for “creative or technical professionals.” And 44% viewed AI PCs primarily as “a gimmick or futuristic technology.”
The survey found that the average user spends a cumulative 899 minutes, or nearly 15 hours, on routine computer-related tasks each week. The reason AI PC users spend longer on tasks than traditional PC users is “because they spend a lot of time figuring out how to best communicate with AI tools to get the answers or responses they want.” “Companies offering AI-enabled products must provide more training to truly demonstrate the potential of ‘everyday AI,’” Intel argued.
According to a 2023 study by Antropic, leveraging AI tools to handle repetitive tasks, streamline workflows, or support research can significantly increase productivity.
Among respondents who were not familiar with AI PCs, only 32% said they would consider purchasing one for their next upgrade, but for respondents who had previously used AI PCs, that percentage rose significantly to 64%. AI PCs can also save users about 240 minutes a week on routine tasks, according to “early data” from multiple studies.
The problem is that many AI PC users aren’t aware of the benefits of AI or don’t know how to access the tools. Antropic said, “Despite AI PCs becoming more available to people, 86% of respondents have never heard of or used AI PCs. Meanwhile, respondents who already own AI PCs actually spend more time on digital daily tasks than those who use traditional PCs,” he said. “More user training is needed to bridge the gap between the promise and reality of AI PCs.” “It is necessary,” he concluded.
Mike Crosby, general manager of industrial advisory services firm Circana, says AI PCs can be a promising solution in a B2B environment. Just three of the 20 federally defined U.S. business sectors, including science, finance and healthcare, account for nearly 50% of all AI PC sales, Crosby said. “Companies must carefully evaluate these new technologies to understand the benefits of innovation and the existing environment. “We are comparing the risks,” he explained.
As 60-70% of the Windows installed base is still using Windows 10, the upcoming end of Windows 10 support in October 2025 will further amplify the demand for AI PCs. Although Microsoft’s ESU provides a temporary reprieve, Crosby expects PC upgrades to progress quickly as the deadline approaches.
editor@itworld.co.kr
Source: www.itworld.co.kr
