AI Is Eating Memory Chips and Phones and Laptops Cost More: Q3 Smartphone Shipments Fall 22% as People Keep Old Devices Longer
All week, we have been writing about AI agents, chatbots and AI search. The side effect few people notice is that every AI model runs in data centers that need enormous amounts of memory, and that race is pushing up the price of phones and laptops. Data from October 2026 shows buyers already feel it: devices cost more, and many people are choosing to keep their old devices longer.
This is not only a concern for people about to upgrade. When the devices in your users' hands are older and slower, every website and app has to run well on them too.
1. AI is eating memory chips, and component prices are soaring
Chipmakers are shifting capacity to high-bandwidth memory (HBM) for AI servers, leaving less for phones and computers. The effect flows straight through to retail prices.
Memory costs up about 350%: according to IDC, memory costs for smartphones have risen by about 350% year over year and are still climbing, although the pace has slowed.
Full-year forecast: Gartner expects DRAM and SSD prices to rise about 130% by the end of 2026, pushing PC prices up 17% and smartphone prices up 13% compared with 2025.
Budget devices hit hardest: Gartner says the sub-$500 PC segment could disappear by 2028, and entry-level phones are hit harder than premium models.
2. Buyers are waiting and keeping devices longer
Rising prices gave the smartphone market its steepest quarterly decline on record in Q3 2026. People are not giving up their phones; they are putting off upgrades.
Shipments down 22%: according to preliminary IDC data released on October 14, 2026, global smartphone shipments fell 22% in Q3 to 255.3 million units. IDC does not expect a recovery before 2028.
6 in 10 have noticed higher prices: a Counterpoint survey of more than 5,000 people in India, China, the US, the UK and Germany found that 63% of those planning to buy a phone in the next six months have seen prices go up, and about half are waiting for deals.
More increases ahead: Counterpoint expects prices of existing models to rise by 20% on average by the end of 2026, with new launches up to 30% more expensive than a year earlier.
Devices kept longer: Gartner expects consumers to keep their PCs about 20% longer by the end of 2026.
The Agenorix view: design for the device your users actually hold
When many people keep their phones for 3 to 4 years, an older mid-range phone is the typical device, not the latest model in the ads. When we build websites and AI assistants for clients, we always start from that reality:
Test on older devices: open your website on a mid-range phone a few years old, not just on the design team's laptop.
Keep pages light: compress images, trim unnecessary JavaScript and avoid loading what readers don't need yet, so pages open quickly even on low-memory phones.
Run AI on the server: an AI assistant should do its processing server-side and display through the browser, so users don't need a new device to use it. The assistant in the bottom-right corner of agenorix.com works this way.
Put the web before the app: don't force people to install another app when their storage is already full. A good mobile website is enough for most needs.
You can see how we apply these principles in Agenorix projects.
Where to start?
If you are planning to upgrade, check your battery and free storage first: a battery replacement and a storage cleanup often add one to two years of use at a much lower cost. If you run a website, open it on the oldest phone on your team and note the 3 slowest spots. If you need a light, fast website with an AI assistant that works well on any device, the Agenorix team is ready to help, and your first consultation is free: get in touch.
Sources: IDC (Oct 14, 2026), Counterpoint Research (Oct 2026), Gartner (Feb 26, 2026)
Share