🔍 Read the full analysis: AI's Growing Appetite For Memory Could Reshape Tech Budgets Until 2030 on ThorstenMeyerAI.com
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
SK hynix says memory demand is likely to outstrip supply through 2030, citing AI infrastructure growth and the time needed to expand production. The forecast comes as memory prices rise and manufacturers face pressure to pass costs on through more expensive devices or lower specifications. The duration and consumer impact remain uncertain, and suppliers’ forecasts come from companies benefiting from higher prices.
SK hynix said October 28 it expects memory demand to exceed supply through 2030, a forecast that could keep pressure on costs for phones, PCs and other electronics as AI data centers buy more chips. The South Korean memory maker cited the long lead time for new factories and the difficulty of producing leading-edge memory; its prediction is a company outlook, not a guarantee of how prices or consumer device sales will develop.
In its quarterly results release, SK hynix said supply was not keeping pace with market growth and forecast that the imbalance would persist through 2030. On the earnings call, Chief Financial Officer Kim Woo-Hyun said rising memory costs and shortages were expected to reduce shipments of PCs and mobile devices, while growth in demand for AI infrastructure would more than offset that decline. The company did not give a specific estimate of how much retail device prices would change.
The source report says a 32-gigabyte DDR5-6000 kit cost at least $447 in the United States on October 29, compared with a recorded low of $72, according to Tom’s Hardware’s price index updated October 26. TrendForce reported contract-price increases of 93% to 98% for ordinary memory chips in the first quarter of 2026. The report’s estimate that prices had reached roughly four times their end-2025 level combines later quarterly increases and forecasts; it is the author’s calculation, not a single reported market index reading.
SK hynix also said demand for high-bandwidth memory, or HBM, from customers was more than twice what it could supply at the time. The company reported long-term agreements with 14 major customers covering more than half of its production capacity for 2027 and 2028. Those arrangements indicate strong contracted demand, but do not establish how much memory will be available to other buyers or how prices will change.
Infographic 01 · Where the memory goes
AI data centers get first call on memory.
High-bandwidth memory (HBM) for AI chips and the memory in your phone come out of the same factories. HBM needs far more wafer per gigabit, and its buyers sign years ahead.
The bakery problem
HBM share of wafers vs. share of the memory bits they produce
Same factories, same wafers: every wafer used for AI memory is missing from phone and PC memory.
Who gets first call
Long contracts lock up supply years ahead
- SK hynix: deals with 14 major customers, usually five years, with deposits; more than 50% of 2027–2028 capacity
- Micron: more than 75% of fiscal 2027 shipments already committed
- HBM demand: customers want “more than double” what SK hynix can supply today
- Cloud budgets: memory ≈ 47% of big cloud providers’ capex in 2026, 68% in 2027 (TrendForce estimates)
Three makers, three horizons, one customer’s view
- SK hynixThrough 2030Demand above supply (company forecast)
- SamsungInto 2028Tight supply, based on market research
- MicronNo “line of sight”Tighter in 2027 and 2028 than in 2026
- Acer (customer)Easing from 2H 2027Says suppliers talk up the shortage
HBM is a wedding cake, phone memory is bread. Same ovens. The cake customers pay more and order years ahead.Thorsten Meyer · my analogy
Sources: TrendForce, 2 Jun (wafer and bit shares, estimates); TrendForce, 25 Aug (capex shares, estimates); SK hynix, 28 Oct 2026; SK hynix call, 28 Oct; Micron call, 30 Sep; The Register, 1 Oct; Samsung call, 29 Oct; Acer via Tom's Hardware / DigiTimes. Wafer-per-gigabit ratio: my arithmetic. All horizons are forecasts.
Powered by Thorsten Meyer AI
How AI Demand Reaches Device Buyers
AI server demand can affect consumer electronics because HBM and conventional memory rely on overlapping production resources. HBM is stacked DRAM designed to move data quickly beside AI processors; conventional DRAM serves as working memory in PCs and phones, while NAND flash stores files and apps. According to the source report’s reading of TrendForce estimates, HBM production uses substantially more wafer capacity per bit than standard memory.
When manufacturers allocate more capacity to HBM, less may be available for other memory products. That does not automatically mean every device will become more expensive: companies may absorb some costs, reduce memory or storage specifications, or change product mixes. But sustained higher component costs can affect both the price and configuration of devices households and businesses buy, while adding budget pressure for manufacturers.
The source material cites IDC figures showing third-quarter 2026 PC shipments down 20.1% and smartphone shipments down 22.0%, as well as an IDC forecast that the average smartphone would cost $581 in 2026, up 27.6%. These figures are reported in the supplied article; they do not by themselves establish that memory shortages caused the shipment declines or forecast price rise.
Products Worth Considering
Amazon Kindle Scribe Colorsoft 64GB (1st Generation)
Samsung Galaxy Tab S10 FE 128GB Wi-Fi Android Tablet, Blue
LG 34GX90SA-W 34-inch Ultragear WQHD OLED Curved Gaming Monitor
As an affiliate, we earn on qualifying purchases.
Why HBM Competes for Wafer Capacity
Memory manufacturers produce several products from related manufacturing capacity. HBM is made by stacking DRAM chips and connecting them vertically, allowing high data throughput for AI processors. Standard DRAM is used as a device’s working memory, while NAND flash provides longer-term storage. The products differ, but the source report says they draw on shared wafer resources, creating a trade-off when manufacturers decide what to produce.
SK hynix’s forecast is not the only supplier outlook cited. Samsung has said tight supply could last into 2028, while Micron has said it has no “line of sight” to a balance between supply and demand. Their statements point to a constrained market, but each is a company forecast and may differ in scope and assumptions. SK hynix’s own financial results also provide context: the source reports third-quarter revenue of 104.6 trillion won and an operating margin of 80%, meaning the supplier benefits from the strong market it describes.
The forecast has also drawn skepticism from buyers. Acer has accused memory suppliers of talking up shortages to protect margins, according to the source report. That allegation has not been independently established in the material provided; suppliers and their customers have different commercial interests in how supply and pricing are described.
“Supply is not keeping pace with the growth of the market.”
— SK hynix, in its October 28 earnings release
How Long Shortages Will Last
2030 is SK hynix’s forecast, not a confirmed end date for shortages. The outlook depends on future AI investment, memory demand, factory construction and how quickly manufacturers can raise output. The supplied material does not give a complete, independently verified forecast for supply and prices across every type of memory.
The effects on retail devices are also uncertain. Some manufacturers may raise prices, while others may reduce specifications or absorb higher component costs. The reported shipment and price figures do not isolate the contribution of memory from other market factors. Nor does the material settle Acer’s allegation about supplier statements; the companies’ financial interests are relevant when evaluating forecasts, but do not by themselves prove that a forecast is inaccurate.
Watch Supply and Device Pricing
Investors and device buyers will be watching upcoming earnings calls and market updates from SK hynix, Samsung and Micron for changes to their supply forecasts, production plans and customer commitments. Further reporting from market researchers such as TrendForce and shipment estimates from IDC may help show whether prices and device volumes continue to move as the source material describes.
The clearest test will be whether new production capacity comes online quickly enough to meet demand, and whether AI infrastructure orders continue to grow at the pace suppliers expect. Until those developments are clearer, SK hynix’s through-2030 outlook should be treated as a company projection rather than a settled timeline for device prices or availability.
Key Questions
What did SK hynix forecast?
SK hynix forecast on October 28 that memory demand would exceed supply through 2030. It attributed the outlook to strong demand and the time and technical difficulty involved in expanding production.
Why does AI demand affect memory for PCs and phones?
AI processors use high-bandwidth memory, which is made from DRAM and competes for production resources with conventional memory. The source report says HBM requires substantially more wafer capacity per bit, potentially limiting capacity available for other products.
Does the forecast mean phones and PCs will definitely cost more?
No. Higher memory costs can put upward pressure on device prices, but manufacturers may also absorb costs or change specifications. SK hynix’s forecast does not specify how much retail prices will change.
Are other memory makers expecting shortages to continue?
The source material says Samsung expects tight supply into 2028 and Micron has said it has no “line of sight” to balance. These are separate company outlooks, not guarantees that supply will remain tight on the same timeline.
What remains uncertain about the shortage?
The duration and severity of any shortage will depend on demand, factory expansion and production decisions. The supplied material also does not establish how much of recent device-price or shipment changes can be attributed specifically to memory costs.
Source: ThorstenMeyerAI.com
