Apple has warned investors that memory chip prices are rising at an “unprecedented” rate. The main reason is competition from AI datacentres for the same supply, and the cost is now landing on phone makers and their customers.
Key takeaways
- Apple has told investors that memory chip prices are rising at a pace it called unprecedented, and that the increase will squeeze its profit margins.
- The Guardian reports that Apple is competing with datacentre operators for memory supplies, which ties consumer electronics costs to demand for AI infrastructure.
- Apple said its decision to raise prices has offset part of the impact, and its revenues came in stronger than expected.
- How long the pressure lasts depends on whether memory manufacturers add capacity faster than AI datacentre demand grows.
Memory is now priced for AI buyers rather than consumers
For most of the smartphone era, memory was a commodity bought in huge volumes by the companies that make phones, laptops and personal computers. Prices went up and down as manufacturers added factories and then waited for demand to catch up. Consumer device makers were among the largest and most influential buyers, and their buying cycles helped decide where prices went.
This article argues that the balance has shifted. Large operators building datacentres for artificial intelligence now compete directly for memory, and they can pay a lot for it because AI hardware is valuable and scarce. When one of the best-resourced consumer electronics companies in the world tells investors that memory costs are hurting its margins, it is a strong sign that consumer device makers no longer set the price. They pay whatever the AI buildout leaves behind.
The consequences reach well beyond one company’s accounts. If memory is priced for AI buyers, every phone, tablet and laptop carries a cost that reflects datacentre spending. Manufacturers can absorb that cost and accept thinner margins, pass it on through higher retail prices, or cut the amount of memory in their products. Apple appears to have done some of the first two.
A large buyer has publicly called the rise unprecedented
The first piece of evidence is the warning itself. According to the Guardian, Apple told investors that memory chip prices were climbing at an unprecedented rate. It said the higher prices would eat into its profit margins, and described itself as competing with datacentres for supplies.
That statement matters because of who made it. Apple buys components in very large volumes and is generally seen as having strong negotiating power with suppliers. Buyers of that size can usually secure long-term contracts, priority allocation and favourable terms. If a buyer in that position still finds memory costs rising sharply enough to warn shareholders, the pressure is probably not coming from ordinary market swings. It points to a structural change in who is bidding for the supply.
The word “unprecedented” is also unusual. Companies tend to describe input costs carefully in investor communications, because the language can move share prices. The Guardian reports that Apple’s shares fell in after-hours trading after the announcement, which suggests investors took the warning seriously. The source does not say exactly how much prices have risen or how big the expected hit to margins is.
Memory makers have strong reasons to favour datacentre customers
The second piece of evidence is the economics of memory production. Only a small number of companies make memory chips at scale, and building new fabrication capacity takes years and very large capital investment. Total supply therefore cannot respond quickly when demand jumps.
AI systems need a lot of memory. Training and running large models means moving huge amounts of data between processors and memory, and specialised high-bandwidth memory has become a key part of AI accelerators. Producing these advanced types of memory uses manufacturing capacity that could otherwise make conventional chips for phones and computers.
When a manufacturer chooses what to make with limited capacity, it will tend to favour the products and customers with the highest margins. Datacentre buyers building AI infrastructure have shown they will spend heavily to get hardware quickly. If production shifts towards those buyers, less is left for conventional memory, and the price of what remains rises for everyone else, including phone makers. The exact capacity split between product types at the major manufacturers is not public in detail, but the direction of the incentive is clear.
This also explains why the pressure falls across the consumer electronics industry rather than on one firm. Apple is the company that made the warning, but other makers of phones, laptops and consoles buy from the same small group of suppliers and face the same competition for capacity.
Apple’s response shows the cost reaching consumers
The third piece of evidence is how Apple says it has responded. According to the Guardian, Apple said its decision to raise prices had offset some of the impact of more expensive memory. That indicates at least part of the higher component cost is being passed to customers.
This matters for the argument because it shows the effect is not limited to supplier contracts and corporate margins. When a company with Apple’s pricing power raises retail prices partly because of component costs, AI infrastructure demand starts to show up in what consumers pay for everyday devices. The source does not say which products changed in price or by how much.
The Guardian also reports that Apple’s revenues were stronger than expected. That fits the picture of a company that can protect its sales while passing on some costs, even as margins come under pressure. Smaller manufacturers with less brand loyalty and thinner margins may have fewer options. Where Apple can raise prices and keep revenue up, others may have to absorb the cost or cut specifications.
The case against: a cyclical squeeze, not a lasting shift
The strongest objection is that memory has always been one of the most cyclical parts of the semiconductor industry. Periods of shortage and soaring prices have repeatedly been followed by periods of oversupply and collapsing prices. On this view, today’s squeeze is the top of another cycle, not a permanent change in who controls pricing.
There is a reasonable logic to this. High prices give memory manufacturers a strong incentive to build more capacity. Once new factories come online, supply could rise faster than demand, especially if AI infrastructure spending slows, pauses or becomes more efficient. Prices would then fall, and consumer device makers could regain their old influence. Earlier booms in the industry ended that way.
A second objection is that AI hardware demand may be less durable than it looks. Spending on AI datacentres depends on investors and companies continuing to expect returns from AI products. If those expectations weaken, datacentre construction could slow quickly, and memory capacity aimed at AI buyers would come back to the wider market.
A third objection is that Apple’s warning may partly reflect its own circumstances. Factors such as product mix, timing of supply contracts or currency movements can affect margins in ways unrelated to a broad market shift. The Guardian’s report does not give a full breakdown of what is driving Apple’s margin outlook, so attributing all the pressure to datacentre competition goes beyond what is known.
Finally, Apple’s results were not weak overall. Revenues were stronger than expected, which shows that rising memory costs, while real, have not yet seriously damaged its business. Calling this a fundamental change could overstate what is so far a margin issue for a highly profitable company.
These objections are serious. Memory’s history of booms and busts is the best reason to doubt that the present situation will last indefinitely.
Evidence that would change the conclusion
Several developments would weaken the argument that AI datacentres now set the price of consumer memory.
A sustained fall in memory prices while AI datacentre construction keeps growing would show that supply can meet both markets and that consumer makers are not simply paying what datacentres leave behind. The same would apply if major memory manufacturers announced capacity expansions large enough to end shortages within a short period.
Investor communications from Apple and other device makers would also be informative. If later reports describe memory costs as stabilising or easing, the current warning would look more like a temporary peak. If other device makers say they have not faced similar pressure, that would point to factors specific to Apple.
A clear slowdown in AI infrastructure spending would test the argument directly. If memory prices fell sharply after datacentre spending eased, that would confirm AI demand as the driver but would also show its effect to be temporary rather than structural.
On the other side, the conclusion would be strengthened if several consumer electronics makers reported the same pressure, if retail device prices kept rising across the industry, or if manufacturers began shipping products with less memory than before. Clear evidence that memory producers are permanently moving capacity towards AI-focused products would point the same way.
Sources and further reading
- The Guardian, technology and business coverage: reporting on Apple’s quarterly results and its warning about memory chip prices.
- Apple investor relations materials: the company’s official results releases and statements to shareholders.
- Semiconductor industry analysis from trade publications: background on memory manufacturing capacity and pricing cycles.
- Annual reports and investor communications from major memory manufacturers: context on production priorities and capacity investment.
Surfaced from the rss:guardian_tech signal “memory chip price rise”. AI-assisted draft, editorially reviewed.
