The AI PC revolution has caused something of an upheaval across the world as organizations seek the upper hand in productivity and efficiency.
As AI workloads demand more compute and power, many businesses are looking for workstations that deliver the advantage without significantly upping their energy bills.
Shown off at Mobile World Congress 2025 (MWC), Intel’s latest product portfolio promises significantly improved performance for AI workloads, while slashing power consumption, increasing security, and simplifying deployment and management.
Boosting performance
For 2025, Intel has rolled out a new line of Intel Core Ultra processors, with the 200V series being unveiled at CES 2025, and additional 200 series processors for laptops, performance notebooks, and desktop & workstation devices announced at MWC – all of which offer significant increases in performance over previous generation processors.
Most notable are the performance metrics for both the 200H and the 200HX workstation processors.
“We’re seeing up to 63% gen over gen gain, which is unusual. Typically, we’re seeing 10-15% gen over gen, so this is really massive,” David Feng, VP/GM, PC Client Segments at Intel told TechRadar Pro at MWC 2025.
To add some additional color to this workload performance, the Intel Core Ultra 9 285HX workstation processor provides a 53% boost in performance in AI and machine learning, a 63% increase in financial services, and a 50% increase in life sciences workloads over the previous generation.
Outside of performance, Intel Wi-Fi 7 can boost connectivity by 5 times, and Intel Bluetooth 5.4 can drop power consumption by 50% for connectivity use cases.
The latest line of Intel Core Ultra processors have been released at a critical time when many organizations are looking to refresh their workstations, especially with the looming Windows 10 end of life arriving later in 2025.
As a result, the Intel Ultra Core 200 series will be available in over 150 commercial designs across more than 10 manufacturers.
Intel vPro
In a welcome move for enterprise, Intel also unveiled a series of updates for the vPro platform across security, productivity, manageability and deployment.
“Security starts with Intel, and we’re very proud of that,” Jen Larson, GM, Commercial Client Segments at Intel said, highlighting that Intel finds around 96-97% of security issues in their products proactively. “We’re breaking what we build,” she adds.
In addition, vPro had 150 mitigation against MITRE ATT&CK framework TTPs, and, when mapped with Windows 11, Secure-core PC, and Microsoft Defender, vPro had 90 software plus hardware mitigations – exceeding the Microsoft SecuredCore PC level 3 standards.
There have also been significant steps forward in making vPro activation easier, cutting the deployment steps from 24 to 6, and removing the need for an on-prem server.
“You can activate in minutes,” Larson said, adding that Intel was able to cut down the deployment time by offering a self-hosted activation method for some of the biggest tools many companies already use, through AWS, at no additional cost to customers.
This new activation method isn’t just for the devices either. “Anybody that has V Pro devices in their fleet can do this,” Larson adds.
In what stands as one of the biggest announcements to come out of Intel’s latest vPro offerings, Intel is also offering a new Assured Supply Chain (ASC) service for select HP and Lenovo mobile and desktop workstations that house select Intel Core Ultra Series 2 chips.
“Our customers can actually attest to that supply chain and where it was manufactured,” Larson adds, explaining that the ASC will provide an additional layer of trust and security for government agencies and businesses working in a world where the global semiconductor supply chain is constantly under threat.
The future of AI PCs
When asked about future developments for AI PCs, Feng says, “We believe in a few years we won’t need to call it an AI PC anymore, because AI will be harnessed as part of pretty much everything we do, through most of the applications we use,” adding that, “no matter who you are, what you do, what form factor you choose – 200 ISVs, 400 plus AI features, more than 500 AI models reported – and that’s all happening more or less last year. You can imagine it’s not a linear curve. It’s going to go up exponentially.”
Feng also points out that this in turn introduces new challenges, such as bigger workloads drawing more power in a time when businesses are looking for greater energy efficiency and battery life.
“The best part actually, is we don’t see any point in the near future where the compute is going to be too big or more than enough,” Feng adds.
“Any level of performance we can deliver on an endpoint device can be easily saturated by just pulling a fraction of the size of the model from the cloud to client,” he says, pointing out the future opportunities in orchestrating a seamless cloud-to-client experience to provide additional AI benefits to businesses.
https://cdn.mos.cms.futurecdn.net/iB9S7M99zU8fs39G6FZdG7-1200-80.jpg
Source link
benedict.collins@futurenet.com (Benedict Collins)
