Samsung's Memory Crisis: How AI Demand is Driving Chip Shortages Through 2028

TL;DR

• Samsung says memory shortages are worsening as AI data centers absorb more HBM and DRAM capacity, and it now sees tight supply lasting into 2028.
• The crunch is already pushing up chip prices for servers and could spill into PCs, smartphones, and other consumer devices as mainstream memory becomes harder to source.
• Analysts and suppliers say demand is being driven by AI infrastructure, with some customers already reserving supply well into 2027.

Samsung is warning that the global memory market is entering a prolonged shortage cycle, with supply pressure expected to intensify in 2027 and remain tight through 2028. The root cause is a dramatic shift in manufacturing priorities toward high-bandwidth memory, or HBM, which is essential for AI data centers and commands far higher margins than conventional chips.

AI data centers are reshaping memory supply

The memory crunch is being driven by one of the biggest demand shifts in semiconductors in years: the race to build AI infrastructure. Modern AI systems require huge amounts of fast memory to keep GPUs and accelerators fed with data, and that has pushed chipmakers to prioritize HBM over standard DRAM and NAND products.

Samsung, SK hynix, and Micron have all been steering more capacity toward AI-focused products, which has reduced availability for the kinds of memory used in servers, PCs, smartphones, and other mainstream devices. Samsung’s own memory chief said “significant shortages” were expected to continue through at least 2027, and some customers have reportedly already secured allocations through that year.

Prices are already rising sharply

The shortage is no longer a theoretical concern. Samsung has already raised prices for several memory modules, and industry reporting points to steep increases across the board. Reuters reported that Samsung lifted prices on some memory chips by as much as 60% compared with September levels, while 32GB DDR5 modules rose from $149 to $239.

Other reports show the broader market moving in the same direction. Contract pricing for DDR5 has more than doubled in some cases, and analysts expect further increases as supply remains constrained. Gartner has forecast a 47% rise in DRAM prices in 2026 because of undersupply in both newer and legacy memory markets.

What this means for consumers and enterprise buyers

For enterprise customers, the impact is immediate: higher memory prices raise the cost of servers, cloud infrastructure, and data-center expansions. That creates a difficult environment for companies scaling AI systems, because memory is not a peripheral component but a core input for performance and capacity.

Consumers may also feel the effects, even if indirectly. As manufacturers redirect production toward AI servers, there is less conventional DRAM available for laptops, desktops, and smartphones, which can lead to higher retail prices or reduced margin flexibility for device makers. In other words, a shortage that begins in the server market can eventually show up in everyday electronics pricing.

Why this shortage could last longer than past cycles

Several factors suggest this is not a short-lived inventory problem. Samsung said unmet demand from this year is likely to roll into next year, making 2027 tighter than 2026. It also warned that there would be little chance of a meaningful increase in incremental supply through 2028.

That longer horizon reflects the scale of AI buildouts now underway. Analysts say demand for both HBM and conventional DRAM is rising as AI workloads expand beyond model training into inference and agentic AI applications, which require persistent memory-intensive computing. That broader demand base makes it harder for memory suppliers to catch up quickly, even if they invest aggressively in new capacity.

Samsung’s profits are soaring, but customers are paying the price

The shortage is helping Samsung’s bottom line. Reuters reported that Samsung was likely headed for another record quarter, with operating profit expected to jump about 18-fold on the back of surging AI memory demand. Bloomberg reporting also described quarterly profit surging 19-fold, reflecting the pricing power that comes with constrained supply.

But the same dynamic is squeezing buyers. Memory makers are benefiting from scarcity, while cloud providers, enterprise IT teams, PC makers, and smartphone manufacturers are facing higher component costs and less room to negotiate. The result is a classic supply-demand imbalance: AI growth is creating a lucrative boom for chipmakers while making memory more expensive for everyone else.

What to watch next

The key question is whether memory suppliers can expand capacity fast enough to ease the crunch before 2028. For now, Samsung’s message suggests the answer is no. With AI data centers still scaling rapidly and HBM remaining the highest-value use for fabrication lines, the mainstream memory market appears set for another stretch of tight supply and elevated prices.

If that trend holds, the most visible effects may not be limited to data-center budgets. They could reach all the way to consumer electronics shelves, where higher component costs are often passed along through product pricing, lower discounts, or slower feature upgrades.

AndroGuider Team

Articles written by the AndroGuider team. We try to make them thorough and informational while being easy to read.