Budget-constrained segments, including government procurement and education, experienced a slightly steeper decline than the commercial market. Both are expected to remain under pressure through the end of the year, although Omdia believes they could see a meaningful recovery in 2027.
The text and the above graph expect PC purchasing to shoot back up in 2027. I don’t buy it.
Purchasing on PCs is down because prices on memory are up. There are only two ways that prices on memory come back down in 2027: a lot more supply enters the market, or demand drops significantly.
There are going to be new fabs coming online in 2027, but they won’t be running at full bore until 2028. A lot of unexpected supply seems unlikely to suddenly materialize in 2027.
I am also pretty skeptical that everyone is going to suddenly stop purchasing memory for AI purposes in 2027. If one wants to bet on a huge retraction in AI-related demand, that might do it. But I think that if that’s what this was based on, they’d be pretty explicit about it.
If purchasers in 2027 suddenly decide “the hell with it, I’m willing to pay a lot more for memory than I was last year”, if price elasticity of demand suddenly falls off, that’d do it. But I don’t see a reason for that to happen. Heck, even if it did, I’m not sure how many other memory consumers there are to try to outbid. The market cannot increase production in that timeframe, so the only way for PC purchasers to use more money is to outbid someone else, to get some memory that would have gone to someone else. It sounds like the big AI companies mostly have contracts in place, so you probably can’t outbid them in 2027, even if you rolled up with more money. I guess maybe you could try to outbid cell phone users for available memory or something.
I’ve also seen a number of articles predicting that memory prices in 2027 will, if anything, be somewhat higher than in 2026.
I think that PC purchases are going to stay low in both 2026 and 2027.
Honestly I feel strangely optimistic after seeing this graph. If the current pricing hellscape only makes the market dip by 14% in 2026 that’s not too bad. And let’s talk absolute numbers: they sold nearly 30 million prebuilts to consumers in 2025??? And only US consumers at that. And for consumers it supposedly only dipped a few hundred thousand? I may just be naive, but this looks like a fairly stable market ( rampocalypse notwithstanding ). I was expecting much worse after hearing the report about ms windows’ base shrinking by the hundreds of millions a year due to mobile.
Year to year growth. It’s axis is on the right side. The bar is taller than the previous year than you had growth, but the precentage of growth was lower than the previous year so the dot is lower on the axis up still above 0%. If it has a red dot it means the year saw not growth, but loss.
The text and the above graph expect PC purchasing to shoot back up in 2027. I don’t buy it.
Purchasing on PCs is down because prices on memory are up. There are only two ways that prices on memory come back down in 2027: a lot more supply enters the market, or demand drops significantly.
There are going to be new fabs coming online in 2027, but they won’t be running at full bore until 2028. A lot of unexpected supply seems unlikely to suddenly materialize in 2027.
I am also pretty skeptical that everyone is going to suddenly stop purchasing memory for AI purposes in 2027. If one wants to bet on a huge retraction in AI-related demand, that might do it. But I think that if that’s what this was based on, they’d be pretty explicit about it.
If purchasers in 2027 suddenly decide “the hell with it, I’m willing to pay a lot more for memory than I was last year”, if price elasticity of demand suddenly falls off, that’d do it. But I don’t see a reason for that to happen. Heck, even if it did, I’m not sure how many other memory consumers there are to try to outbid. The market cannot increase production in that timeframe, so the only way for PC purchasers to use more money is to outbid someone else, to get some memory that would have gone to someone else. It sounds like the big AI companies mostly have contracts in place, so you probably can’t outbid them in 2027, even if you rolled up with more money. I guess maybe you could try to outbid cell phone users for available memory or something.
I’ve also seen a number of articles predicting that memory prices in 2027 will, if anything, be somewhat higher than in 2026.
I think that PC purchases are going to stay low in both 2026 and 2027.
Honestly I feel strangely optimistic after seeing this graph. If the current pricing hellscape only makes the market dip by 14% in 2026 that’s not too bad. And let’s talk absolute numbers: they sold nearly 30 million prebuilts to consumers in 2025??? And only US consumers at that. And for consumers it supposedly only dipped a few hundred thousand? I may just be naive, but this looks like a fairly stable market ( rampocalypse notwithstanding ). I was expecting much worse after hearing the report about ms windows’ base shrinking by the hundreds of millions a year due to mobile.
What is the line on that graph supposed to represent? There’s years where the bars are higher but the line goes down
Year to year growth. It’s axis is on the right side. The bar is taller than the previous year than you had growth, but the precentage of growth was lower than the previous year so the dot is lower on the axis up still above 0%. If it has a red dot it means the year saw not growth, but loss.