🔍 Read the full analysis: Why The AI Boom Could Keep Memory In Short Supply For Years on ThorstenMeyerAI.com
Get the little things that make your day delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
SK hynix said on October 28 that memory demand is likely to exceed supply through 2030, citing factory construction times and the difficulty of advanced manufacturing. AI data centers are drawing capacity toward high-bandwidth memory, while suppliers and market analysts report rising prices and pressure on PC and phone costs. The duration and scale of shortages remain uncertain, and suppliers’ forecasts are disputed by some customers.
SK hynix expects memory demand to exceed supply through 2030, the South Korean chipmaker said October 28, warning that new factories take years to build and advanced manufacturing is difficult. The outlook adds to concern that AI data-center demand will keep pressure on memory prices and availability for PCs and phones, although the precise length and severity of any shortage are not settled.
SK hynix said in its third-quarter earnings release that memory supply is not keeping pace with market growth. The company cited the technical difficulty of leading-edge production and the long lead time for new facilities as reasons it expects the supply-demand imbalance to continue through 2030. That is a company forecast, not a guarantee that every memory product will be scarce until that date.
On the earnings call, Chief Financial Officer Kim Woo-Hyun said higher memory costs and supply shortages were expected to reduce shipments of PCs and mobile devices. He added that growth in demand for AI infrastructure was expected to more than offset that decline. SK hynix also said customers’ requested volume of high-bandwidth memory, or HBM, was more than twice what the company could currently supply. Long-term agreements with 14 major customers cover more than half of its production capacity for 2027 and 2028, according to the company.
Other suppliers have offered their own outlooks: Samsung has said tight supply could continue into 2028, while Micron has said it has no “line of sight” to a supply-demand balance. Those statements are not identical forecasts, but they point to continued strain. Market data cited by ThorstenMeyerAI.com says contract prices for conventional memory rose sharply in the first quarter of 2026 and continued to increase in subsequent quarters; the article’s longer-term price comparison partly relies on forecasts.
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
AI Orders Compete for Memory Capacity
The pressure reflects a production trade-off, not simply a surge in demand for more memory sticks. HBM is a form of DRAM built by stacking chips for fast data transfer beside AI processors. It uses production resources also needed for conventional DRAM used in computers and phones. Market researcher TrendForce estimated that, by the end of 2026, HBM would take about 22% of wafer starts among the three largest makers while accounting for about 9% of memory bits produced. For the end of 2027, its estimates were about 30% of wafer starts and 13% of bits.
That imbalance can make manufacturers favor HBM, which is in high demand from AI infrastructure buyers, over some consumer-market products. If chip costs rise or supplies are limited, device makers may pass on costs, delay products, or adjust specifications. The source article reports that some manufacturers have reduced storage, camera or memory specifications in mainstream devices, and that some phone makers have reused recycled memory chips. Those measures vary by product and company; they do not establish that all new devices will have lower specifications.
For consumers, the concern is that memory prices could affect the cost and configuration of everyday electronics, not just servers. A cited Tom’s Hardware price index put a 32-gigabyte DDR5-6000 kit at a minimum of $447 in the United States, compared with a recorded low of $72. That is a comparison between the stated current price and historical low, not a measure of how much prices will rise from here.
Products Worth Considering
BowFlex Results Series SelectTech 552 Adjustable Dumbbell
Fitbit Versa Smart Watch – Black Aluminum, One Size (S & L Bands Included)
Oura Ring 4 Smart Ring – Silver – Size 8
As an affiliate, we earn on qualifying purchases.
Memory has different functions in devices. DRAM is short-term working memory used while a computer or phone runs; NAND flash stores files and apps when the device is off. HBM is a specialized type of DRAM, stacked to move large volumes of data quickly for AI processors. Standard DDR5 memory in PCs and LPDDR memory in phones serve different product needs, but they depend on manufacturing capacity that suppliers must allocate among markets.
SK hynix’s forecast follows a period of strong demand and high earnings for the company. Its third-quarter release reported revenue of 104.6 trillion won and an operating margin of 80%, according to the source article. Samsung’s results the following day showed a different picture across its businesses: its memory division set a record while its phone and network business lost money. The results underline how memory supply conditions can benefit chip producers even as downstream device businesses face cost pressures.
Reported shipment and price figures also need care in interpretation. The source cites IDC estimates of a 20.1% fall in PC shipments and a 22.0% fall in smartphone shipments in the third quarter, alongside a projected average smartphone price of $581 for 2026, up 27.6%. The supplied material does not provide enough detail to treat those figures as direct proof that memory costs alone caused the changes.
“Supply is not keeping pace with the growth of the market.”
— SK hynix, in its October 28 earnings release
Forecasts Face Market and Customer Questions
The 2030 timeline is SK hynix’s expectation, and actual supply will depend on factory construction, manufacturing yields, customer orders and the pace of AI investment. Samsung’s stated outlook runs into 2028, while Micron has described itself as unable to see when supply and demand will balance. None of those statements establishes exactly when prices will peak or when consumer-device availability will improve.
There is also disagreement over how much of the tight market reflects physical constraints and how much reflects suppliers’ business choices. Acer’s accusation that suppliers are overstating scarcity is a claim, not an independently established finding. The companies projecting a prolonged shortage are also reporting strong memory results, giving them a commercial interest in the market outlook. The supplied material does not include a detailed response from suppliers to Acer’s allegation.
Price and shipment measures cover different products and periods, and cannot by themselves show that memory costs caused every increase in device prices or every drop in shipments. The source material also does not establish how quickly additional capacity will come online or how much of it will be allocated to HBM rather than conventional memory.
Capacity Plans and Prices to Watch
The next indicators will be suppliers’ quarterly updates on factory investment, production capacity, HBM output and customer contracts. SK hynix’s agreements for 2027 and 2028 suggest that a substantial portion of its planned output is already committed, but the company has not said that all capacity is sold or how allocations will change if demand shifts.
Device makers’ pricing and product specifications will show how much of the pressure reaches consumers. Buyers and manufacturers will also be watching whether AI infrastructure investment continues at the pace suppliers expect. Until more capacity is producing and market demand becomes clearer, the timing of any easing remains uncertain; the available forecasts point to continued tightness, but differ on its likely duration.
Key Questions
What did SK hynix forecast?
SK hynix said it expects memory demand to exceed supply through 2030, citing the time needed to build factories and the difficulty of advanced manufacturing. It is the company’s forecast, not a certainty that every type of memory will remain scarce for that entire period.
Why is AI demand affecting PC and phone memory?
AI processors use high-bandwidth memory, a specialized form of DRAM. HBM and conventional memory rely on overlapping production resources, so allocating more capacity to HBM can leave less available for some PC and phone products.
Will phones and PCs definitely become more expensive?
No. Higher memory costs can put pressure on device prices or specifications, but the effect will vary by manufacturer and product. The supplied figures do not show that memory costs alone determine device prices.
Do all suppliers expect shortages through 2030?
No. SK hynix forecast that demand would exceed supply through 2030. Samsung has described tight supply into 2028, and Micron has said it has no “line of sight” to balance. These are different company statements, not a single agreed industry timeline.
What could make the outlook change?
New production capacity, manufacturing yields, changes in AI investment and customer demand could all affect the balance. The available information does not specify how quickly those factors will shift supply or when prices might ease.
Source: ThorstenMeyerAI.com
