What Omdia says has changed
Omdia has been tracking AI's effect on the hardware market. It has seen messaging move away from offering AI as a destination. The new position pitches AI as a platform that can support more applications and services on a PC.
Kieren Jessop, a principal analyst at Omdia, made the case at the Canalys Channel Forum in Barcelona. He said there had been a gap between what customers expected and the products they ended up using. Computer Weekly's Microscope reported his remarks. The forum ran from 6 to 8 October 2026.
Jessop's diagnosis is that the earlier pitch was too narrow. It treated the AI PC as something with relevant features straight out of the box. In his view, most businesses want PCs that are secure, cost-effective, long-lasting and easy to manage. He described the vendors' response to the "lacklustre reception" of the past couple of years. They now want the AI PC seen as something to build services, software and applications on.
The report has no shipment figures, survey results or other numbers behind the "lacklustre" verdict. It is an analyst's assessment, not a measured outcome. It also doesn't say AI PCs have failed everywhere, or that vendors have dropped their out-of-box AI features.
The keyboard isn't going anywhere
Jessop also tempered expectations about how people will use PCs. He doesn't expect everyday interaction to change fundamentally over at least the next couple of years. Voice might play a slightly bigger role, but he expects the keyboard to remain what most people use daily.
That is a forecast, not a usage study. It does explain why vendors may be cooling on the idea that AI will visibly reshape the desktop experience.
Windows 11 is still the stronger upgrade argument
Last year Jessop argued that Windows 11 motivated upgrades more than AI did. That theme has remained. The industry still finds it hard to justify a premium for fresh hardware.
His 2025 comments were reported by Computer Weekly at the time. The majority of PCs being sold were AI-enabled, but that functionality was not the key consideration in upgrading. He said the commercial refresh had been holding up the market, and that Windows 11 refresh explained the growth. He also said most consumers and businesses want a similar laptop experience: performance, battery life, extra commercial features and compatibility with existing applications.
Other coverage points the same way, though it comes from different sources and should be read as separate evidence. Windows Latest reported in July 2026 that Windows 11 PC sales were largely driven by Windows 10's end of life, not AI PCs. It also noted that OEMs such as HP still insist that AI PCs are driving Windows 11 sales. The industry is not unanimous, and vendor framing is worth reading with that in mind.
Platform promise versus buyer reality
For IT buyers, the shift sets a long-term promise against immediate needs. The promise is that capable hardware will support future AI software and services. The needs are the four Jessop named: security, cost, longevity and manageability.
The report doesn't give a procurement checklist. As my own synthesis from Jessop's stated priorities, a sensible lens for an AI-PC premium would ask:
- Security: Does the device improve your security posture today, or only on a roadmap?
- Cost: Is the premium justified by software you actually run, not software that is promised?
- Lifespan: Will the hardware stay useful across a normal refresh cycle, and does the AI capability extend that or just add price?
- Manageability: Can your existing management tooling handle the new features without extra overhead?
There are open questions the report doesn't answer. It names no future applications, gives no arrival dates and doesn't say whether they will need dedicated on-device AI hardware. It also doesn't say whether the platform framing will produce stronger demand. Treat the platform pitch as a hypothesis, not a result.
Earlier forecasts, for context
Omdia's earlier Canalys insight offers a reference point. It defined an AI-capable PC as a desktop or notebook with a dedicated chipset or block to run on-device AI workloads. That is not a new definition for 2026, and the forum report doesn't say Omdia has revised it.
The same insight forecast that 19% of PCs shipped in 2024 would be AI-capable, rising to 60% in 2027. It also expected businesses to receive 60% of AI-capable shipments by 2027. Those are old projections, not verified outcomes. A high share of AI-capable shipments says little about whether buyers bought for AI.
The insight also named challenges: getting channel partners on board, overcoming user resistance, and creating meaningful on-device experiences. A platform-and-services pitch is arguably a response to the last of these. Nothing in the reporting shows it has been solved.
The channel dilemma
The forum also split on hardware inventory. Omdia fellow Steve Brazier urged resellers to fill their warehouses. Jay McBain countered that current demand was fuelled by AI infrastructure spending and was a "moment in time".
Jessop took a middle path. Resellers shouldn't stock up and pay warehousing fees while waiting for customers to buy. Hardware will still become more important, so they should keep watching the market. This is attributed advice, not a stock-level forecast, and the report doesn't establish what is driving demand beyond McBain's view.
Bottom line
- The AI PC story is shifting from features available now to a platform for later.
- Jessop's reading is that businesses still buy on security, cost, longevity and manageability.
- Windows 11 remained a stronger refresh driver than AI in the reported analysis.
- Whether the platform pitch pays off depends on software and services the report doesn't identify.
If you're planning a refresh, ask vendors which specific applications justify the premium and when they will ship.