Samsung’s Profits Are Soaring. Why Is It Warning of a Worse Memory-Chip Shortage?

How can record profits and a shortage happen together?

The component making Samsung richer could also make your next phone more expensive.

Samsung Electronics reported quarterly operating profit of 107.4 trillion won, approximately $80 billion, for July–September. Alongside those earnings came a warning: AI-driven demand for memory could exceed industry supply by a wider margin in 2027.

There is no economic mystery here. When customers compete for scarce components, suppliers can command higher prices. Selling more valuable products can also improve their earnings.

For manufacturers buying those components, the same conditions create a problem. They must absorb higher costs, raise prices or change what they put into their products.

Samsung’s results capture both sides of that transaction.

WHAT EXACTLY IS RUNNING SHORT?

“Chips” is too broad a description. The immediate concern is memory and storage used across AI systems, conventional servers, computers and phones.

A processor performs calculations. Memory keeps the information it needs readily accessible. Storage holds information for longer.

DRAM is working memory. NAND flash provides storage in products such as solid-state drives. High-bandwidth memory, or HBM, is a specialised form of DRAM designed to move large amounts of data rapidly.

HBM stacks memory dies vertically and connects them through pathways running through the silicon. That structure helps supply data to demanding AI processors. Samsung’s technical documentation describes both the stacked construction and the manufacturing challenges involved.

These products serve different purposes. A shortage does not mean every chip is equally scarce or that a phone contains the same memory package as an AI accelerator.

WHY DOES AI NEED SO MUCH MEMORY?

A powerful processor needs a steady supply of data to remain useful.

Training an AI model involves processing enormous quantities of information. Running a trained model also requires access to its parameters and the data being processed for users.

Memory capacity determines how much information can remain close at hand. Bandwidth determines how quickly it can move.

Adding computing power without sufficient memory can leave expensive processors waiting for data. That makes memory a central part of the AI infrastructure race.

The demand also extends beyond HBM. Samsung identifies server DRAM and solid-state storage among the products it is prioritising as AI infrastructure expands.

The pressure therefore spreads across several parts of the memory industry.

WHY CAN’T MANUFACTURERS SIMPLY MAKE MORE?

They are expanding. The difficulty is how quickly additional usable supply can arrive.

Semiconductor production requires specialised facilities, equipment and tightly controlled processes. Advanced memory adds challenges involving stacking, connections, heat and the proportion of manufactured products that meet specifications.

Competitor SK hynix says the technical difficulty of advanced processes and the long lead times for new facilities are preventing supply from keeping pace. Its latest results forecast demand exceeding supply through 2030. That is the company’s outlook, rather than a guaranteed timetable.

Capacity decisions also involve trade-offs. Reuters reported in January that Samsung and SK hynix were prioritising more lucrative AI-related products while warning of tighter supplies for computer and smartphone manufacturers.

A surge in investment cannot instantly produce finished, tested chips.

IS THIS JUST A TEMPORARY BUYING RUSH?

Manufacturers are describing demand that stretches beyond the current quarter.

SK hynix says long-term agreements already cover more than half its production capacity for 2027 and 2028. Such commitments give suppliers greater visibility when planning investment.

They also show why smaller customers may struggle to secure supply against buyers willing to commit far ahead.

Nevertheless, orders and forecasts do not eliminate uncertainty. AI customers could revise spending plans. New production could arrive faster than expected. More efficient systems could alter how much memory particular workloads require.

The shortage outlook rests on assumptions about both demand and supply. Those assumptions need continued scrutiny.

DOES EVERY PART OF SAMSUNG BENEFIT?

Its own accounts show the limits of that assumption.

Samsung’s semiconductor division reported operating profit of 106.5 trillion won. Its mobile and networks businesses together recorded an operating loss of 0.6 trillion won, with the company citing cost pressures among the factors affecting earnings.

That does not establish memory prices as the sole cause of the loss. It does show why the consolidated profit figure cannot describe every business within the group.

A component supplier benefits from strong selling prices. A finished-device business must still persuade customers to buy the resulting product at a viable price.

Even within one company, those economics can pull in different directions.

WHAT DOES THIS MEAN FOR INDIAN BUYERS?

The pressure is already visible in India’s smartphone market.

Counterpoint Research says shipments fell 20% year-on-year in July–September 2026. Shipments below ₹20,000 declined 47%. Its assessment identifies a memory-led cost squeeze alongside softer consumer demand.

Those figures need careful interpretation. Shipments measure devices entering distribution channels, rather than purchases directly recorded at every shop. A shrinking price segment can also reflect models moving into higher price bands.

Still, higher component costs leave manufacturers with difficult choices. They can accept lower margins, increase prices, reduce discounts or alter configurations.

For a buyer working within a fixed budget, that could mean paying more for a comparable device or receiving less memory and storage at a familiar price.

The outcome will vary by brand, inventory, supply contracts and currency movements. Samsung’s warning does not establish a uniform price increase for every phone or laptop.

DO SAMSUNG’S PROFITS PROVE THE AI BOOM IS SUSTAINABLE?

They demonstrate that supplying memory to the infrastructure buildout is currently lucrative.

They cannot establish whether every company buying that infrastructure will earn an adequate return.

A supplier receives revenue when it sells a component. Its customer must subsequently turn the equipment into services that generate enough income to justify the investment and operating costs.

Those are different stages of the business case.

WHAT SHOULD WE WATCH NEXT?

The important developments are whether new capacity becomes productive, whether memory prices keep rising and whether customers maintain their spending commitments.

Consumer demand deserves equal attention. If higher device prices cause people to delay upgrades, the pressure can spread beyond the original component shortage.

Samsung’s record quarter shows how valuable scarce memory has become. Its warning suggests that buyers of that memory, including companies making everyday devices, may face a harder negotiation next year.