🔍 Read the full analysis: From Data Centers To Shopping Baskets: AI's Memory Impact on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- 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 exceed supply through 2030, as AI data centers compete for manufacturing capacity used to make memory for PCs and phones. The forecast comes as memory prices rise and device shipments fall, although the companies benefiting from tight supply are also the source of the shortage outlook.
SK hynix expects memory demand to exceed supply through 2030, the South Korean chipmaker said October 28, warning that higher memory costs and supply constraints could reduce PC and mobile-device shipments. The outlook links the expansion of AI data centers to rising costs for chips used in everyday devices, including phones, laptops and desktop PCs.
In its third-quarter release, SK hynix said supply was not keeping pace with market growth. It attributed the outlook to the technical difficulty of advanced manufacturing and the long lead time needed to build new factories. The company’s statement forecasts a supply shortfall through 2030; it does not say that particular consumer devices will remain unavailable or that their prices will rise by a fixed amount.
The company’s finance chief, Kim Woo-Hyun, made the consumer-device connection on the earnings call. He said PC and mobile-device shipments were expected to decline because of the cost burden from higher memory prices and supply shortages, while AI infrastructure demand was expected to more than offset that decline. The statement is a company forecast, not a reported count of future shipments.
SK hynix said demand for high-bandwidth memory (HBM), the stacked memory used alongside AI processors, was more than twice the quantity it could supply at the time. It also said long-term agreements with 14 major customers covered more than half of its production capacity for 2027 and 2028. The company described those contracts as generally five years long and said they include deposits.
There are signs of pressure on both prices and device sales in the source material. A 32-gigabyte DDR5-6000 kit was listed at at least $447 in the United States, compared with a lowest recorded price of $72 cited by the source from Tom’s Hardware. TrendForce reported that contract prices for conventional memory rose 93% to 98% in the first quarter of 2026. Those figures concern different measures and time frames; they should not be read as a direct price forecast for every retail product.
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 Memory Demand Reaches Consumers
AI companies are buying large quantities of HBM for data-center processors. HBM is made from DRAM, the same broad memory category used in computers and phones, and the source says it draws on manufacturing resources also used for conventional memory. When manufacturers allocate more production to HBM, less capacity may be available for other memory products, adding pressure to supply and prices.
That pressure can affect consumers even when they are not buying AI hardware. Memory is a component in PCs and phones, so higher input costs can contribute to more expensive devices, reduced specifications or delayed upgrades. The source cites an IDC forecast for an average smartphone price of $581 in 2026, up 27.6%, and reports a steep fall in third-quarter shipments. These figures do not, by themselves, establish that memory costs alone caused the price increase or shipment declines.
The impact also reaches manufacturers and retailers that must decide whether to absorb higher component costs, pass them on, or adjust product designs. For buyers, the practical issue is not only the price of a separate memory kit: it may also be how much storage or memory is included in a new device and what the device costs overall.
Products Worth Considering
Apple iMac 2024 All-in-One Desktop with M4 Chip, 24‑inch Retina Display, 16GB Unified Memory, 256GB SSD, Silver
Dell XPS 13 9350 Laptop, 13.4 FHD+ Display, Intel Ultra 7 256V, 16GB DDR5, 1TB SSD
Apple MacBook Pro 2023 with M2 Pro Chip, 14‑inch, 16 GB RAM, 512 GB SSD – Space Gray
As an affiliate, we earn on qualifying purchases.
Why HBM Competes for Factory Capacity
DRAM temporarily holds data a device is using, while NAND flash stores files and apps after power is off. HBM is a form of DRAM built by stacking memory chips and connecting them to move data quickly beside an AI processor. Conventional PC memory, such as DDR5, and phone memory are made in different product formats, but they depend on overlapping semiconductor production resources.
TrendForce estimates cited in the source say HBM will account for about 22% of wafers started by the three largest memory makers by the end of 2026, while representing about 9% of memory bits produced. For the end of 2027, its estimates are about 30% of wafers and 13% of bits. The source author calculates from those estimates that HBM uses roughly three times as much wafer capacity per gigabit as other memory. That ratio is the author’s calculation, not a direct quotation from TrendForce.
SK hynix is not the only company signaling tight supply. The source says Samsung expects constraints to last into 2028, while Micron has said it has no line of sight to a balanced market. These are separate company assessments with different stated time frames; SK hynix’s forecast extends furthest.
“Supply is not keeping pace with the growth of the market.”
— SK hynix, in its October 28 earnings release
Forecasts, Prices and Competing Interests
The duration and severity of the shortage remain forecasts, not settled outcomes. New factories, manufacturing yields, customer orders and changes in AI investment could affect how much memory becomes available and when. The source does not provide a complete forward price series for consumer memory or a measure isolating memory’s contribution to phone and PC prices.
The companies forecasting tight supply also stand to benefit from strong prices. The source reports that SK hynix posted record quarterly revenue and an 80% operating margin. It also says Acer accused suppliers of talking up shortages to protect margins. That is an allegation attributed to Acer, not an established finding; the material does not include a response from the memory makers or independent evidence resolving the dispute.
Samsung’s reported results add another sign of uneven effects: its phone and network business lost money while its memory business set a record, according to the source. The available material does not provide enough detail to attribute that result solely to memory costs or to quantify how much consumers have paid because of the supply squeeze.
Signals to Watch Through 2028
The next evidence will come from quarterly company updates on production, customer commitments, memory prices and device demand. SK hynix’s long-term customer contracts cover more than half of its production capacity for 2027 and 2028, according to the company, making actual deliveries and capacity additions relevant measures of whether supply catches up.
Consumers and device makers will also get a clearer picture as PC and smartphone shipment data and retail prices are reported in coming quarters. The present forecasts point to continued pressure, but they do not establish exactly when prices will ease or how manufacturers will balance memory capacity between AI hardware and consumer products.
Key Questions
What did SK hynix forecast?
SK hynix said memory demand is expected to exceed supply through 2030, citing manufacturing complexity and the time required to build new production facilities.
Why can AI data centers affect phone and PC memory?
AI processors use high-bandwidth memory, which is made from DRAM and draws on manufacturing resources also used for other memory products. Greater HBM production can leave less capacity for conventional memory, although the exact effect on consumer supply can vary.
Does the forecast mean phone and PC prices will keep rising through 2030?
No. SK hynix forecast a prolonged gap between memory demand and supply. The source does not establish a specific consumer-device price path through 2030, and device prices depend on factors beyond memory costs.
Are the reported memory price increases the same as retail device price increases?
No. The cited TrendForce figures concern contract prices for memory chips, while the $447 figure is a retail listing for a memory kit. Neither figure directly measures the full cost of a phone or PC.
Source: ThorstenMeyerAI.com
