If you were hoping the RAM and SSD “memory apocalypse” was near the end, Jefferies Equity Research just poured cold water on that plan. As reported by Wccftech, Jefferies expects memory prices to rise 40 to 50% in Q3 2026, then another 30 to 40% in Q4 2026. And it does not stop there. The research projects a 40 to 45% year-on-year increase in 2027.

So why does this matter beyond the gaming PC forums? Because your cost baseline for everything from laptops to app hosting is tied to the price of memory. The report frames the current pressure as driven by AI data centers buying up available stock, tightening supply in a way that hits both consumer builds and product roadmaps. If RAM and SSD costs keep stepping up through 2026 and into 2027, device makers have two options: eat the margin damage or raise prices. Either choice shows up in budgets, demand, and planning.

Jefferies is only predicting a recovery thanks to increased production in 2028, and even then the rebound is limited: around 15 to 20%. That matters because markets often expect “recovery” to mean fast normalization, but here it sounds more like a partial thaw, not a full reset. In other words, the new normal may stay elevated for longer than many buyers and even many internal planning models assume. That’s the quiet risk for CFOs and procurement teams: you can’t just wait out the cycle and hope you’re back to baseline by next quarter.

There is also a narrative clash about where the relief is supposed to come from. Former Samsung boss Kyung-Hyeon Kye had predicted an earlier end to the memory crisis by pointing to an increase in Chinese manufacturing. But the source notes that, now, it seems like Chinese memory is not actually selling for less. That detail flips a common assumption on its head. If additional production does not translate into cheaper market pricing, then procurement savings may not arrive on schedule even when capacity expands. For boards and executives, that is a reminder that production volume and consumer pricing do not always move in sync when supply chains, contracts, and demand priorities are distorted.

Zoom out and the incentives get clearer. AI data centers are the demand-side megaphone in this story, and when they buy available stock aggressively, it can crowd out other buyers, including consumer electronics channels. The source says the memory apocalypse has mainly been a downer for anyone looking to build a new PC or pick up a Steam Machine, which is basically a shorthand for: gaming hardware gets hit twice, once on availability and again on pricing.

And the pressure is not limited to PCs. The source points out that Apple recently announced it is jacking up the price of iPads and Macbooks by hundreds of dollars, and suggests iPhones likely will not be far behind. That is not “evidence” of memory alone causing everything, but it does reinforce the broader mechanism: component and supply constraints can flow through to consumer pricing. If memory costs keep rising while major device makers raise prices, it compounds affordability stress for everyday buyers.

Second-order, the source connects this to a cultural and economic tension: whether ordinary folks still think it is worth building data centers to make coding easier for some programmers when the tradeoff is that people cannot afford new phones. That phrasing is provocative, but the underlying idea is straightforward. If AI infrastructure spending stays hot while hardware costs rise, the benefits and costs may land unevenly across different consumer categories. That can influence public sentiment, political scrutiny, and future adoption curves.

Finally, the story hints at a potential inflection: the AI bubble cannot burst soon enough, and it notes that OpenAI lost $38.53 billion in 2025. The reason to care, from an executive perspective, is that demand-side momentum often drives pricing power. If AI spending slows, memory buying could cool. But even the presence of losses does not guarantee a rapid slowdown, and Jefferies is still forecasting price increases deep into 2027. So decision-makers should treat this as a multi-year pricing risk, not a short-term volatility event.

Bottom line: Jefferies expects memory prices to keep climbing through Q4 2026, then rise 40 to 45% year-on-year in 2027, before only partial relief of roughly 15 to 20% production gains in 2028. For anyone planning hardware launches, cloud margins, or procurement budgets, that is a long runway for cost pressure. Plan for the “memory apocalypse” to be less of an apocalypse ending and more of a slower, more expensive reality that your models have to survive.