AI-capable PC adoption to rise significantly by 2027

AI-capable PCs are set to transform the user experience and usher in a new era of innovation, with more than 60 per cent of all PCs shipped using the technology by 2027.

Those were some of the key findings from research carried out by analysts Canalys, who surmised that integrating dedicated AI acceleration hardware into PCs can enable innovation in efficiency, productivity, collaboration and creativity.

Canalys concluded that to capitalise on the new technology, which is set to become mainstream in the coming years, vendors must prioritise ecosystem collaboration to develop differentiated solutions, communicate benefits clearly to customers, and uphold strong privacy and ethical standards in using data.

Canalys said that initially AI-capable PCs will be based on today’s narrow hardware requirements, with dedicated chipsets or blocks needed to run AI workloads on-device. But it added that the criteria will need to be broadened to grade products on overall AI experience as technical capabilities, use cases and customer requirements evolve.

As dedicated AI blocks become mainstream, Canalys said that additional considerations around hardware benchmarks, pre-installed models and use case performance will be required to properly distinguish AI capabilities. Thus, an adaptable approach to defining and segmenting AI-capable PCs will be needed as the market matures.

In 2024, almost one in five PCs shipped will be AI-capable, with that figure growing rapidly to more than 60 per cent by 2027, according to Canalys. Of that total, businesses will account for 60 per cent of shipments by 2027.

That growth will be driven by the upcoming end of support for Windows 10 prompting a major refresh cycle from 2024 to 2025, said Canalys, presenting an opportunity to migrate users to new AI-capable devices. Commercial deployment will spike as the benefits become apparent, resulting in more than 100 million shipments in 2025.

Supporting this, Canalys said that major Windows PC OEMs will be launching waves of new AI-accelerated models, starting in 2024, aligned with next generation CPUs and the transition to Windows 11. So far, HP, Dell, Lenovo, Acer and Asus have all announced plans to deliver new AI-capable PCs, timed alongside Intel and AMD’s new CPU product roadmaps.

By bringing such offerings to market during the 2024/25 Windows refresh, Canalys said that this gives OEMs an opportunity to accelerate upgrades. But their success depends on ecosystem collaboration to communicate the benefits of their products clearly and to deliver differentiated out-of-the-box experiences.

Concurrently, Canalys said that leading chip vendors are incorporating dedicated AI acceleration blocks – NPUs - into CPUs to significantly improve on-device AI capabilities. Processors integrating NPUs from Intel, AMD, Apple and Qualcomm will be used to power these new AI-capable PCs.

But true differentiation will come from co-engineering hardware and software with OEMs and other partners, said Canalys. Such initiatives, like Intel’s partnership with more than 100 ISVs, demonstrates the importance of collaboration to balance performance, energy efficiency and usability to deliver meaningful user benefits.

To fully capture the opportunity, however, Canalys said that vendors must clearly communicate all of the benefits of AI-capable PCs to both users and partners. They must also adhere to high privacy and ethical standards around using data, and foster ecosystem collaboration to deliver impactful, differentiated experiences.

Canalys concluded that the value of AI-capable PCs is not yet clear to many businesses. Therefore, proof points showing quantified outcomes from its deployment and use are needed.

And as on-device AI requires access to user data, Canalys said that maintaining high privacy and ethics standards is key to building trust. This includes transparency, accountability and user control. Only responsible development and communication will accelerate adoption and result in AI being integrated more ambiently across devices. 

Posted under: