With the first wave of AI PCs hitting the market soon, the question now is who the PC makers will sell these devices to. Meanwhile, Intel already has the answer to who will lead the AI PC market in the early launch phase. "The business side of the world will be the immediate beneficiaries because that's where people are looking for instant productivity gains, and they will be the ones to take it up first.” George Chacko, Director of Global Account Sales, APJ at Intel, explained why corporates will lap up AI PCs for productivity and efficiency benefits. Chacko, however, mentioned that consumer adoption of AI PCs may take some time even as corporations will be quick to computers optimized for artificial intelligence-heavy workloads. “There are concerns around security, data privacy, and usage, and therefore, there might be a few steps of correction along the way, but the uptake will be much faster,” Chacko told indianexpress.com on the sidelines of a Lenovo event in Bangkok, Thailand. Intel, alongside Microsoft and PC vendors such as Lenovo and others, has been talking about a new type of personal computer powered by new chips like Intel’s Core Ultra with dedicated “Neural Processor Units” (NPUs) for months. These PCs, also referred to as “AI PCs,” have AI capabilities that can help accelerate AI tasks locally, rather than reaching out to cloud servers (like ChatGPT and Microsoft Copilot). “When you are doing it on your PC, you are doing it locally. That’s not only faster but also more efficient, and you can also have a very localised version of your local language model, for example, that's very specific to enterprise usage, that you would run off. I think there are advantages, and I believe the future is to be done locally,” he said. For years, artificial intelligence has been used to improve PC tasks, mostly behind the scenes, such as video upscaling, microphone noise reduction, and power/battery management. However, in the new era of generative artificial intelligence, Large Language Models (LLMs) can generate human-like responses to users’ queries through simple prompts, using applications that are designed to optimise AI at the foundational level. But it’s important to realise that when you pull up ChatGPT, use Copilot on Windows 11, or interact with a popular AI chatbot, the processing isn’t done on the computer. Instead, all that processing is happening in a data centre somewhere, which requires a lot of resources to run an AI model. You can perform Gen AI tasks using the Stable Diffusion text-to-image model or the LLAMA large language model locally with reduced or no cloud support. However, to achieve the desired results, you will still need a PC with a faster CPU as well as GPU to run those on-device generative artificial intelligence applications. Chacko said Intel wants to make mainstream PCs—the ones powered by its Core Ultra processor containing a special neural processing engine, or NPU—that can run these genAI models locally. So, from the hardware point of view, we are getting there, even if the current “AI PCs” may not fully meet Microsoft's performance requirements. However, despite the AI branding on PCs and a dedicated NPU, the software isn’t fully ready to take advantage of “AI PCs” yet. Microsoft may be pushing a new Copilot key as part of the AI PC push, but Windows 11 doesn’t do anything with an NPU that’s inside those AI PCs. Until Windows get an AI makeover and applications like Copilot run locally on PCs, the dream of AI PCs may not fully materialise. Chacko says Intel’s AI PC acceleration program, launched in October, will help speed up the fine-tuning of applications and software to run on its AI PCs. The idea behind the acceleration programme is to get software developers interested in AI-powered features for their apps, and it’s now being expanded to reach smaller developers through an AI PC developer programme. Intel is also targeting hardware makers with the AI PC acceleration program, helping them to optimise and enable their hardware for Intel’s AI PCs. Intel’s on-device strategy and the Gen AI makeover of PCs are its efforts to respond to the fast-changing deployment of AI across core products and services. The company is targeting 100 million AI PCs by 2025 as the tech industry looks for the next growth driver to boost sales of personal computers which have fallen in the past few quarters. But the competition is heating up, with Qualcomm planning to launch new Snapdragon X Elite-powered laptops that could offer a significant performance leap for Windows on Arm. In fact, the release of an optimised version of Google’s Chrome browser for Windows on Arm this week promises improvements in the experience for Chrome users on Qualcomm-powered Windows laptops set to be available in the second half of the year. In the coming months, both Apple and Microsoft are expected to announce their AI strategies in detail at their respective developer conferences, where we could see an in-depth look at software and possibly new AI hardware. “Everyone is looking for a killer app in the consumer space, and we say the killer app is the choice,” Chacko responded when asked which application will drive sales of AI PCs. “One software would appeal to you, another one would appeal to me, the third one to a content creator, and the fourth one to someone who's doing photo editing. So for us, the killer app is a choice.”
