AI chips, smart software engines now power our laptops

Early this year, Chinese laptop maker Lenovo unveiled four new gaming laptops with an embedded artificial intelligence (AI)-powered chip called Lenovo LA AI. The branding of the chip, which will be displayed on the company’s Legion Pro 7, 7i, 5 and 5i gaming laptops, is aimed at helping the company deploy its soc-called software--the Lenovo AI Engine+--which will allow the laptop to automatically tune the system's performance using machine learning (ML) technology. Lenovo claims that the laptop is able to adjust performance depending on the user’s activity with its AI/ML-powereed chip and software engine.

Lenovo isn’t the only PC maker to do so. Unlike 2020, when such features were implemented at the software level only, chipmakers including Intel, Advanced Micro Devices (AMD) and Nvidia are now making AI processing a core part of their chips. 

While Apple, too, had built a neural engine into its new M-series chips, Intel, Nvidia and AMD together command a much larger market share in the PC market. Chips from these firms are usually used for Windows-based laptops, which compromise over 75% of the PC market, according to statistics website StatCounter.

For instance, AMD unveiled a chip called the Ryzen 7000 on January 4. This chip comes with an AI engine called Ryzen AI built in, and several PC makers have already committed to using the chip on their devices. Intel too is said to be working on something similar with its Meteor Lake processors, expected to launch later this year.

The moves are akin to Samsung, Qualcomm, Apple and Google, which have built chips with specialized AI functionality into their mobile chips. Google, for instance, builds Tensor chips for its Pixel smartphones to enhance features like the Google Assistant.

This will also enable developers to build more AI-driven programs for laptops, the same way apps use AI functionalities on phones. “Many of the experiences we get on PCs demand trillions of operations per second. AI engines can run them without taxing CPU and GPU while consuming a few hundred milliwatts. It means you get incredible AI, without sacrificing performance or battery life. This means you can run powerful language models without getting in the way of anything else that you might be doing on your PC at the same time,” said Panos Panay, chief product officer at Microsoft, at the CES during the unveiling of the Ryzen AI, last month.

To be sure, device makers have been offering AI-based features on laptops for a couple of years now. HP, for example, offers AI-based privacy alerts, noise reduction, and enhanced video and sound experience on its Spectre X360 laptops. Dedicated chips will enhance such features, on paper. Microsoft’s Windows Studio Effects software also uses AI to blur background during calls, filter noise, and maintain eye contact with the webcams even when the user is looking away. Chips like these will enhance such software, and possibly make them faster.

“Ryzen AI engine is different because it is the world’s first integrated AI engine on an X86 processor. This means that AI workloads before now have been executed using standard processor or graphics silicon that is not optimized to process AI workloads,” said Vinay Sinha, managing director of sales, AMD India.

He explained that just like the invention of the graphics processing unit (GPU), which enhanced graphics performance on PCs, purpose-built AI processing hardware enables new types of approaches to AI workloads. “Now that OEMs have access to this hardware, we anticipate AI-driven software to improve with more features, performance, and battery optimization over time,” he added. 

“AI is the next big phase in laptops, with faster processor cores, intelligent features, and better graphic chips – to cater to the evolving consumer requirements in today’s world,” said Nitish Singal, Head, Personal Systems (Consumer) at HP India. 

Further, experts believe that AI engines will help PC makers market laptops better for enterprise buyers and developers too, even though this may not have an immediate impact on the demand for PCs.

“In the long term, AI in the operating system (OS) and apps creates not just automation benefits, it starts to become predictive and provides active training and learning. This starts to change the skills equation. Now PC providers can market something new, with PCs a Workplace tool,” said Ranjit Atwal, senior director Analyst, at Gartner.

Atwal said that smaller PC brands are typically more innovative with these new tools. “They can create more specialized AI modules that work specifically for different verticals or industries,” he added.

Navkendar Singh, associate vice president, devices research at IDC, noted that the end objective is to make PCs do such things that one cannot do on smartphones. “These features can help ,but I reckon more in creator/developer space and other commercial use cases. From a consumer's point of view, the end-use case needs to be very compelling. Maybe Gaming and use cases coming out of Metaverse can appeal, but those are still a few years away realistically,” he said.