🔍 Read the full analysis: AI And The Memory Shortage: What To Expect For Phones, PCs And IT on ThorstenMeyerAI.com
Get hardware and tech essentials 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 says memory demand is likely to exceed supply through 2030, as AI data centers compete for production capacity with consumer devices. Prices for memory have surged, while industry forecasts point to higher device costs, fewer shipments and possible cuts to specifications. The shortage outlook is the manufacturers’ forecast, and the timing and severity remain uncertain.
SK hynix expects memory demand to exceed supply through 2030, the South Korean chipmaker said on October 28, warning that rising costs and shortages could reduce sales of PCs and mobile devices. The forecast reflects competition for production capacity between AI data centers and consumer electronics, and comes as prices for memory chips and retail PC memory kits have climbed sharply.
SK hynix said supply is not keeping pace with market growth. It attributed the prolonged imbalance to the difficulty of producing advanced memory and the long lead time for new factories. The company’s forecast is its own outlook, not a guarantee that shortages will last for a specific period or affect every product equally.
On the company’s earnings call, Chief Financial Officer Kim Woo-Hyun said rising memory costs and limited supply are expected to weigh on PC and mobile-device shipments. SK hynix also said customer requirements for high-bandwidth memory, or HBM, are more than twice what it can currently supply. Long-term agreements with 14 major customers cover more than half of its production capacity for 2027 and 2028, according to the call.
Consumer-market indicators show the pressure already being felt. The source reports that a 32-gigabyte DDR5-6000 kit cost at least $447 in the United States, compared with a recorded low of $72, according to Tom’s Hardware’s price index updated October 26. TrendForce reported that contract prices for conventional memory chips rose 93% to 98% in the first quarter of 2026. These figures refer to different measures—retail kit pricing and chip contract prices—and should not be treated as the same price series.
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 Devices
AI data centers use HBM, a form of DRAM designed to move data quickly beside AI processors. HBM and the DRAM used in PCs and phones rely on the same broad pool of wafer-manufacturing capacity. Because HBM requires more wafer input per amount of memory produced, manufacturers have a commercial incentive to allocate capacity toward it, where demand and returns are strong. That can leave less capacity for conventional memory used in laptops, desktops and phones.
For consumers and businesses, the effects may appear as higher device prices, fewer available models or lower specifications, rather than as an obvious shortage of chips at retail. The source cites industry reports that some manufacturers are reducing storage, memory or camera specifications to control costs, and that some phone makers are reusing recycled memory chips. Those adjustments vary by product and company; they do not establish that every new device will have reduced specifications.
Businesses face a related planning problem: memory is an input across personal computers, servers and other IT equipment. If prices stay elevated or supply is allocated through long-term contracts, buyers may have less flexibility over timing and configuration. The extent of that impact depends on supplier capacity, product demand and how much cost manufacturers pass on.
Products Worth Considering
Apple MacBook Air 15-inch Laptop with M5 Chip, 16GB Memory, 512GB SSD, and Liquid Retina Display
Apple MacBook Pro 14‑inch (2021) with M1 Pro, 16 GB RAM, 512 GB SSD – Space Gray
Apple MacBook Pro 14 (2021) with M1 Max, 64 GB RAM, 2 TB SSD, Silver
As an affiliate, we earn on qualifying purchases.
Why Memory Capacity Is Tight
DRAM is working memory that devices use while running; NAND flash stores files and applications when a device is powered off. HBM is a specialized form of DRAM made by stacking memory chips and connecting them vertically. It is built for the high data-transfer needs of AI accelerators, rather than as a drop-in replacement for the memory in a phone or PC.
New factories and advanced production processes take time to build and qualify, according to SK hynix. At the same time, AI infrastructure providers are placing large orders and seeking supply well in advance. TrendForce estimates cited in the source say HBM is expected to use a growing share of wafer starts at the three largest memory makers, while producing a smaller share of total memory bits. The source’s estimate that HBM uses roughly three times as much wafer per gigabit is the author’s calculation from those figures, not a direct company statement.
Other suppliers have also described a tight market, but their timelines differ. Samsung has said supply constraints could persist into 2028, while Micron has said it has no “line of sight” to a market balance. Those statements show that concerns extend beyond one company, but forecasts remain subject to changes in investment, demand and production yields.
“Supply is not keeping pace with the growth of the market.”
— SK hynix, in its October 28 earnings release
How Long Will Prices Stay High?
The forecasts do not establish how much memory prices will rise, how long retail prices will remain elevated or which devices will see the largest changes. SK hynix’s outlook depends on demand and production conditions over several years, while Samsung and Micron have offered different descriptions of the expected timeline. New capacity, shifts in AI investment or weaker consumer demand could change the balance.
There is also a dispute over how to interpret the tight market. Memory makers cite constrained supply and strong demand; Acer has accused suppliers of overstating the shortage to protect margins. The source does not provide evidence resolving that allegation. Shipment and pricing figures cited in the source are industry estimates, and the available material does not establish how much of each device-price increase is attributable specifically to memory.
What to Watch in Upcoming Results
Investors, device makers and customers will be watching upcoming earnings reports for changes in memory pricing, factory investment, customer commitments and the pace of AI data-center orders. Suppliers’ updates should clarify whether their forecasts are changing, though company guidance remains a projection rather than a guarantee.
For buyers, the near-term development to track is whether manufacturers pass memory costs through to retail prices, adjust configurations or change product availability. The source does not identify a single date when supply is expected to ease across the market. Until more capacity comes online or demand shifts, the balance between AI-related orders and consumer-device needs remains a central uncertainty.
Key Questions
What did SK hynix forecast?
SK hynix said on October 28 that it expects memory demand to exceed supply through 2030, citing production challenges and the time needed to build new facilities. This is the company’s forecast, not a confirmed end date for shortages.
Why can AI data centers affect phone and PC memory?
AI accelerators use HBM, a specialized type of DRAM. HBM and conventional memory use overlapping wafer-manufacturing capacity, so allocating more capacity to HBM can limit supply for memory used in consumer devices.
Are phone and PC prices certain to rise?
No specific price increase is certain. The source reports substantial increases in memory pricing and cites an industry forecast for a higher average smartphone price, but manufacturers decide how much of their costs to pass on and which product specifications to change.
Do all memory makers expect the shortage to last through 2030?
No. SK hynix forecast tight supply through 2030. Samsung has described constraints lasting into 2028, and Micron has said it has no line of sight to market balance. The companies’ outlooks differ and remain forecasts.
What is still unknown about the shortage?
The duration and severity of supply constraints, their eventual effect on specific device prices and the amount of new production that will come online are not settled. The source also reports a dispute over whether suppliers are overstating tightness, which is not resolved by the material provided.
Source: ThorstenMeyerAI.com
