AI PCs: A New Era of Personal Computers?

AI PCs stand apart from standard laptops by incorporating an additional processor. These AI chips primarily handle the control and management of the device’s processes. They also enhance AI-specific functions, such as personal assistants or graphic processors.AI PC owners can use AI functionalities without relying on cloud services, which also leads to lower battery consumption. 

Although machine learning, particularly with large language models, typically requires substantial cloud computing resources, these new notebooks or laptops do not replace data centers. Instead, they enhance the user's experience by speeding up interactions with chatbots and other AI apps like image generators and voice emulators.

A Paradigm Shift? 

Notably, the last decade has been largely focused on reducing device weight, simplifying user interfaces, and boosting energy efficiency. The rise of cloud computing and the integration of cloud-based AI services fostered a belief that high computational power was unnecessary for personal computers—most users wouldn't fully utilize it, anyway. This shift also diminished the need for large-capacity hard drives and other previously standard features, a concept widely embodied by Chromebook (Google’s laptops that run on ChromeOS), which operate mainly on online apps.

However, this approach eventually led manufacturers into a commercial trap. Since PC performance became dependent not on hardware specifications but on cloud capabilities, the incentive to frequently update devices waned.

Today, the average laptop replacement cycle in the corporate sector is around four years, with individual users updating their devices even less often. According to Counterpoint Research, global PC sales fell by 14% in 2023. Although there was a slight recovery towards the end of the year, sales figures are still well below normal levels.

The solution has emerged in the form of integrating AI directly into personal computers. This shift, however, necessitates a departure from previous development paths: the next generation of PCs will need to enhance computational power and expand local storage capacities to support AI's autonomous functions. This inevitably means moving away from the past focus on lightweight laptops and high energy efficiency. 

The Race Is On

Qualcomm, Intel, AMD, and other tech giants have kicked off marketing campaigns for AI-equipped personal computers.

Intel’s Vice President of Client Computing Group David Feng announced that in 2024 the company aims to deliver AI processors for 40 million PCs, and this number is expected to rise to 60 million by 2025, potentially capturing over 20% of the global PC market. 

Qualcomm has observed a 20% increase in sales of AI PCs featuring their chips within a month of their launch, despite these models being priced 48% higher than non-AI equivalents. Samik Chatterjee, an analyst at JPMorgan Chase, predicts that Qualcomm could hold a quarter of the AI PC market by 2026.

Counterpoint Research has forecasted that if the PC market grows at an annual rate of 3% from 2024 to 2027, the AI computer segment could expand by 59% each year. Such growth could lead to a swift transition from traditional computers to AI PCs, which are expected to eventually dominate the market.

Michael Dell, CEO of Dell Technologies, put it plainly  

Do you want to buy a PC that is not capable of doing those AI things that you’ll want to do in the future? I don’t think so

Beyond the Hype

Despite the enthusiastic announcements, there’s another side to the story. According to research from IDC, only 3% of PCs sold in 2024 will have the computational power necessary to be classified as AI PCs. A market share of 40% is only anticipated by 2027-2028. So, what’s the real issue here?

Experts divide AI PCs into several categories based on the complexity of the tasks they can perform. For instance, the basic AI models currently on the market are not yet equipped for generative artificial intelligence. These models are expected to be phased out within the year, and replaced by AI-advanced laptops with more robust capabilities, eventually leading to models fully equipped to handle GenAI.

Currently, consumers are hesitant to pay extra for simulated AI features, such as eye-tracking during video conferences or enhanced capabilities of built-in GPUs that still depend on cloud services. Why opt for a limited AI on a laptop when full capabilities are accessible in the cloud?   

Additionally, having an AI processor is insufficient; comprehensive software and application adaptation to the new computer capabilities is necessary. This adaptation will also require time. For example, Microsoft plans to introduce GenAI only in Windows 12, slated for release in the latter half of 2025.

While applications like Microsoft Office, WhatsApp, and Photoshop have been optimized, many older applications still used in business have not been updated. This reality potentially limits the adoption of AI PCs in the corporate sector. 

However, transitioning to this new type of PC is simply a matter of time. The overall trend in the market is becoming increasingly apparent. With the broad deployment of PCs equipped with generative AI and built-in Large Language Models (LLMs), users will have the opportunity to exponentially speed up computational processes—by millions of times. The functioning of two GenAIs—one on the PC and the other on a cloud service—will operate in a dialogic mode. It’s expected that cloud services will manage the most demanding processes, while the outcomes will be processed directly on the user’s PC.