🔍 Read the full analysis: How The AI Boom Is Driving A Memory Shortage Through 2030 on ThorstenMeyerAI.com
Get movie nights 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 it expects memory demand to exceed supply through 2030, citing the time needed to build factories and the difficulty of advanced production. AI data centers are increasing demand for high-bandwidth memory made using capacity that also serves other memory products, but the duration and consumer-price effects remain forecasts rather than certainties.
SK hynix said on October 28 that it expects memory demand to exceed supply through 2030, linking the long outlook to rising AI demand and the time required to build production facilities. The forecast matters beyond data centers because high-bandwidth memory for AI systems draws on manufacturing capacity also used for memory in PCs and phones.
In its third-quarter release, the South Korean chipmaker said supply was not keeping pace with market growth. It attributed the gap to the technical difficulty of leading-edge processes and the long lead time for new factories. The release’s 2030 outlook is the company’s forecast, not a guarantee that shortages or high prices will persist unchanged through that year.
On the earnings call, Chief Financial Officer Kim Woo-Hyun said rising memory prices and supply shortages were expected to reduce PC and mobile-device shipments, while growth in AI infrastructure demand would more than offset that decline. The official release did not mention those device shipment effects; they were discussed on the call.
SK hynix also said customer requirements for high-bandwidth memory, or HBM, were more than twice what it could supply at the time. It reported long-term agreements with 14 major customers, generally lasting five years, that cover more than half of its production capacity for 2027 and 2028. Those commitments indicate strong demand, but do not establish how prices or allocations will evolve for other memory products.
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 Memory Competes for Chip Capacity
HBM is a form of DRAM designed to move large volumes of data quickly beside AI processors. It is made by stacking memory chips, and its production uses silicon wafers and manufacturing resources that can also serve conventional DRAM and other memory. When manufacturers assign more capacity to HBM, less may be available for other products unless total production grows enough to compensate.
That competition can affect device makers and consumers through component costs, product availability or specifications. The source reports that a 32-gigabyte DDR5-6000 kit in the United States cost at least $447 on October 29, compared with a recorded low of $72 for that kit type. That is a price snapshot, not evidence that every memory product has followed the same path or that AI alone caused the change.
The stakes extend to manufacturers deciding which products to build and customers deciding when to upgrade. SK hynix’s forecast suggests the capacity squeeze could be a multi-year issue, but its scale will depend on production expansion, demand and how suppliers allocate output.
Popular right now
Bose QuietComfort Wireless Bluetooth Headphones with Active Noise Cancellation, Mic, USB‑C Charging, Deep Bass, 24‑Hour Battery – White Smok
iRobot Roomba Max 715 Pet Robot Vacuum with AutoEmpty Dock
Bose QuietComfort True Wireless Earbuds with Active Noise Cancellation, Black
As an affiliate, we earn on qualifying purchases.
How HBM Uses More Production
Computer memory serves different roles. DRAM temporarily holds data a device is using; NAND flash stores files and apps after power is off. HBM is a specialized form of DRAM, while DDR5 memory sticks and mobile LPDDR serve PCs, servers and phones. They are distinct products, but their production draws on overlapping manufacturing capacity.
TrendForce estimated that by the end of 2026 HBM would account for about 22% of wafer starts at the three largest memory makers while supplying about 9% of memory bits. Its estimates for the end of 2027 were about 30% of wafer starts and 13% of bits. Those figures illustrate why HBM can absorb a larger share of wafer capacity than its share of total bits; they do not mean every manufacturer has the same mix.
Other suppliers have also signaled tight conditions, according to the source material. Samsung has said supply could remain tight into 2028, while Micron has said it has no “line of sight” to a supply-demand balance. These are company assessments with different wording and horizons, not a single industry-wide deadline.
“Supply is not keeping pace with the growth of the market.”
— SK hynix, in its October 28 earnings release
How Long the Shortage Will Last
The forecast through 2030 comes from SK hynix, a supplier that benefits from strong memory demand. The source material also reports that Acer has accused suppliers of talking up shortages to protect margins. That is an allegation; the material does not provide independent evidence that suppliers are overstating the supply outlook. Company forecasts and customer criticism should be treated as interested claims, not settled findings.
The available information does not show how much new production will come online, how quickly manufacturers can raise yields, or how demand for AI systems will change. Nor does it establish a precise retail-price path for PC and phone memory. The cited kit price is a single US market observation, and the supplied material does not provide a complete price series or a comparable baseline for all memory categories.
Shipment figures and projections also require care. The source says PC and smartphone shipments fell in the third quarter and that IDC expects higher average smartphone prices this year, but those trends alone do not isolate memory costs from other factors. The size of memory’s contribution to device prices and sales remains unclear.
Factory Growth and Device Costs
The next evidence will come from quarterly results and updates by SK hynix, Samsung and Micron on production, customer commitments and supply. New factory capacity and progress in advanced manufacturing will help show whether supply can catch up with demand; those changes take time, according to SK hynix.
For buyers, upcoming PC and phone releases may offer a clearer view of how manufacturers respond. Companies could pass on higher component costs, adjust memory configurations or change other specifications. The source material reports such changes in some mainstream products, but it does not establish that all brands or models will follow the same approach.
Until suppliers provide more production detail and market data show how prices move, the 2030 date is best read as SK hynix’s current outlook, not a confirmed end date for a shortage. Future results may narrow or extend the expected period of tight supply.
Key Questions
What did SK hynix forecast?
SK hynix said on October 28 that it expects memory demand to exceed supply through 2030, citing difficult advanced production and the time needed to build factories. This is the company’s forecast.
Why does AI demand affect PC and phone memory?
AI systems use high-bandwidth memory, which draws on manufacturing capacity also used for other memory products. If capacity shifts toward HBM and total output does not rise enough, other memory can become tighter, although the extent varies by product and supplier.
Does the forecast mean all memory prices will keep rising until 2030?
No. SK hynix forecast a supply-demand imbalance, not a specific price path for every memory type. Prices will also depend on production, demand, supplier decisions and other market conditions.
Is a shortage through 2030 confirmed?
No. The 2030 timeframe is SK hynix’s expectation. Other suppliers have described tight supply on different timelines, while the future pace of factory expansion and AI demand remains uncertain.
Source: ThorstenMeyerAI.com
