Apple’s foldable iPhone is also a memory story

Key takeaways

• AI demand is making memory scarcer and more expensive, pushing the boom from data centres into consumer electronics.

• Apple’s launch tests whether strong pricing power can protect profits when the cost of building an iPhone rises.

• South Korea offers AI exposure through memory, but that also means geographical diversification may hide similar economic risks.

Tomorrow, 9 September 2026, Apple takes the stage. Most attention will probably go to folding screens, cameras, thinner devices and artificial intelligence.

But arguably the more interesting investment story sits inside the phone.

One of its most basic ingredients, memory, has suddenly become far more valuable because the same technology is increasingly needed by a completely different customer: the artificial intelligence data centre.

Apple has already warned that rising memory costs and component shortages are affecting its business. Now it is expected to unveil its first foldable iPhone just as those costs are climbing sharply.

The road from Apple’s new iPhone to Nvidia’s AI servers is shorter than it looks.

The chip nobody sees is becoming hard to ignore

Artificial intelligence needs more than powerful processors.

Graphics processing units, or GPUs, perform the calculations. Memory gives those processors fast access to the enormous quantities of data they constantly need. Think of the GPU as the chef and memory as the kitchen counter. A brilliant chef becomes less useful if every ingredient sits in another room.

That is why high-bandwidth memory, a particularly fast type of memory used in AI systems, has become another important bottleneck alongside processors, networking and electricity.

The numbers show how quickly the economics have changed. Global dynamic random-access memory, or DRAM, industry revenue reached almost USD 155 billion in the second quarter of 2026, up about 60% from the previous quarter, according to TrendForce. Samsung Electronics, SK Hynix and Micron dominate the industry.

The important point is not simply that demand is strong. Supply cannot increase overnight.

Memory factories cost billions and take years to build. Producers therefore allocate scarce capacity towards the products offering the best returns. Today, that increasingly means server memory and high-bandwidth memory for AI.

The result is straightforward. More capacity for the data centre can mean less available for smartphones and personal computers.

For investors, this is an important mechanism to understand: when supply cannot respond quickly, strong demand can become pricing power.

Apple now gets part of the AI bill

This is where the AI boom leaves the server room and arrives in your pocket.

TrendForce estimates that memory represented only around 10% of the component cost of Apple’s equivalent premium iPhone a year ago. For the coming generation, it expects that share to rise dramatically, helping push the total component cost of the iPhone 18 Pro about 38% higher than its predecessor.

That creates a simple test for Apple: can innovation offset inflation?

The expected foldable iPhone could give consumers their clearest reason in years to reconsider what an iPhone looks like. But Apple is trying to sell that innovation while the device itself becomes more expensive to build. This is where Apple’s moat becomes practical rather than theoretical.

Its brand, ecosystem and enormous installed base give it more room than most manufacturers to charge premium prices. A weaker company may have to absorb rising component costs and watch margins shrink. Apple has more options: raise prices, adjust specifications, accept slightly lower margins or encourage customers towards higher-value services.

Pricing power is easy to admire when costs are falling. Its real value appears when suppliers send a bigger bill.

Apple is not entering an empty market either. Huawei and Xiaomi launched new foldables on 7 September. Huawei already controls about 68% of China’s foldable market, while Samsung remains the global leader. Xiaomi’s latest foldable reportedly uses memory from Chinese producer CXMT.

So the question is not whether Apple can make a phone that folds. It is whether Apple can turn foldables from a niche category into the next premium smartphone upgrade cycle.

Follow the bottleneck and you arrive in Korea

The supply-chain story eventually leads to South Korea.

Semiconductors accounted for roughly 41% of South Korean exports during the first eight months of 2026, with semiconductor exports rising about 170% from a year earlier. AI memory is increasingly shaping not only individual companies, but the country’s wider economy and stock market.

Samsung sits on both sides of the story as a smartphone manufacturer and a major memory producer. SK Hynix is more directly exposed to the memory boom, while US-listed Micron provides another route into the same cycle.

That creates a useful portfolio lesson. Buying Korean equities may provide geographical diversification from US technology stocks. But if both positions ultimately depend on the same AI investment cycle, the economic diversification may be smaller than the map suggests. Different countries do not always mean different risks.

Risks: memory has a long memory for cycles

Memory remains a cyclical industry. Today’s shortage can eventually encourage too much investment, leading to tomorrow’s excess supply. Slower AI infrastructure spending, rising inventories or falling memory contract prices would be early warning signs that pricing power is weakening.

There is also no guarantee that foldables become mainstream. TrendForce expects them to represent only around 2% of global smartphones in 2026. Apple could successfully launch a beautiful product into a category that remains relatively small.

Finally, higher prices test consumers as much as Apple. Strong pricing power helps, but it does not make demand unlimited.

Investor playbook

• Follow scarcity, not only sales. Ask which components become harder to obtain when a new technology grows quickly.
• Watch margins alongside growth. Apple’s pricing power matters only if it can offset higher costs without damaging demand.
• Look through country labels. Korean and US technology holdings can still depend on the same AI spending cycle.
• Watch the supply response. New factories and rising inventories can eventually turn today’s bottleneck into tomorrow’s oversupply.

The invisible part of tomorrow’s launch

Tomorrow, millions of people will watch Apple unveil what could be the biggest change to the iPhone in years. They will debate the screen, the price and whether consumers really need a phone that folds.

One of the most important components will remain invisible.

AI is turning memory from something consumers rarely consider into one of technology’s critical bottlenecks. That raises costs for Apple, strengthens the economics of Samsung, SK Hynix and Micron, and is reshaping South Korea’s market along the way.

The investing lesson is simple: when a new technology becomes enormous, do not only ask who sells it. Ask what it makes scarce. Sometimes the quieter part of the supply chain has the louder economics.