4 min Devices

There is much uncertainty surrounding the AI PC, but everyone is exploring the options

There is much uncertainty surrounding the AI PC, but everyone is exploring the options

Just about every hardware manufacturer wants to invest in the AI PC. While many opinions exist about integrating the technology into hardware, concrete elaborations are still scarce.

Lenovo said at the Canalys EMEA Forum 2023 that it is working on a new form factor. The manufacturer suspects it to be ready by 2025. This is, however, only a first step toward an actual device. Moreover, Luca Rossi, Senior VP and President of the Intelligent Devices Group at Lenovo, says it is impossible to give updates publicly.

“You will see that there will be some specific views of what an AI PC is, and there will probably even be different levels of AI PC available on the market,” he said at the conference, according to The Register. “An AI PC is a device that is constantly learning about you. It is a PC that models your personal foundation within the PC, and it is a PC that can interact with you in a more natural way,” he added.

Challenges for the AI PC

As with conversations with a chatbot or a home device, the data an AI PC collects about you is stored. Rossi knows this brings some new security challenges. HP’s CEO, Enrique Lores, elaborates on this challenge further at the same conference. According to Lores, data will not be transferred to the cloud, as this storage location will be completely locked out. That means users of an AI PC will also not be tied to a cloud subscription.

Canalys CEO Steve Brazier places yet another piece of the puzzle. Brazier explains that these devices will have a higher price tag and use more power, so the chips must handle their power consumption smarter.

Also read: Laptops are getting more AI features: what is (and will be) possible?

AI has more potential for data centers

Dell, another hardware manufacturer that previously indicated its belief in the PC revolution through AI, now argues it will mainly benefit from servers with AI. To the company’s predictions, it believes it will generate a seven percent increase in sales of AI servers and data storage solutions.

PCs will also drive higher sales, but Dell estimates the increase in this market at 2.5 percent. Dell shared the predictions at a meeting for investors. Jeff Clarke, Dell’s COO, reiterated to investors that primarily Microsoft will fuel AI PC sales. The Windows maker is pushing users to a newer version of the operating system and wants future versions to take advantage of artificial intelligence capabilities. According to Clarke, this is very positive for the PC market, which has declined after COVID-19.

Will Microsoft come up with an AI chip?

It will become more clear how serious Microsoft bets on AI during Ignite 2023, according to rumours. The Information reports based on sources that the company will then show a chip dedicated entirely to artificial intelligence. It would be used in the company’s data centers, and to support AI functions in its productivity suite. According to the sources, however, it would not be a complete replacement for Nvidia’s GPUs, but through this route, the company can make up for shortages at the chip supplier. The rumours are similar to OpenAI’s plans, which were also leaked through sources. Since Microsoft is the main investor in this AI producer, it is not excluded but also not certain that the plans are linked.

In general, just about every vendor is somehow working on the AI PC. But the train of these devices has just left, and the full momentum is not there yet. Manufacturers seem to be able to demonstrate more concretely what artificial intelligence changes for these devices in a couple of years. In the meantime, we do see AI accelerate software development. However, this will be further driven by the fact that users do not have to make significant investments in new products to run this software. Moreover, according to chip manufacturer AMD, the AI applications currently flooding the market can usually be used with pre-existing chips that already support the technology to a small extent.