JupiteX Get the app
Economy & Business17 Nov 2026 · about 7 min

Samsung, SK hynix spend W88tr on capex as cash holdings soar

The brief

The two companies spent 88.16 trillion won on capital expenditure from January through September, according to the article. That is a large investment in the facilities and equipment needed to make chips. It matters because both companies are expanding production amid strong demand for AI-related memory. Samsung spent 56.94 trillion won, up 76.6 percent from the same period a year earlier. About 52.75 trillion won went to its Device Solutions division, which oversees semiconductors. SK hynix spent 31.22 trillion won in nine months, already exceeding its full-year 2025 spending of 30.17 trillion won. It invested 13.62 trillion won in the third quarter alone. The article links the spending to expanded and upgraded chip production facilities, including capacity for high-bandwidth memory and other AI-related memory chips. The figures show both companies committing heavily to meet growing demand. They do not specify how much new output the spending will deliver or when that capacity will come online.

01

How much did Samsung and SK hynix spend on capital investment in the first nine months of the year?

The two companies spent 88.16 trillion won on capital expenditure from January through September, according to the article. That is a large investment in the facilities and equipment needed to make chips. It matters because both companies are expanding production amid strong demand for AI-related memory.

Samsung spent 56.94 trillion won, up 76.6 percent from the same period a year earlier. About 52.75 trillion won went to its Device Solutions division, which oversees semiconductors. SK hynix spent 31.22 trillion won in nine months, already exceeding its full-year 2025 spending of 30.17 trillion won. It invested 13.62 trillion won in the third quarter alone.

The article links the spending to expanded and upgraded chip production facilities, including capacity for high-bandwidth memory and other AI-related memory chips. The figures show both companies committing heavily to meet growing demand. They do not specify how much new output the spending will deliver or when that capacity will come online.

02

What does capital expenditure mean, and what kinds of things can chipmakers buy with it?

Capital expenditure, or capex, is spending on assets a company expects to use over time rather than on routine day-to-day costs. For a chipmaker, it can fund production facilities and upgrades that support making more or different chips. That spending matters because semiconductor manufacturing depends on specialized sites and equipment.

Typical examples include building or expanding a factory and purchasing tools used in chip production, such as machines for processing silicon wafers. These examples draw on general industry knowledge; the article itself says Samsung expanded and upgraded chip production facilities, but does not list specific purchases. Samsung directed 52.75 trillion won of its capital spending to its Device Solutions division, which oversees semiconductors.

The article reports that Samsung and SK hynix spent 88.16 trillion won on capex in the first nine months. SK hynix said it was expanding capacity to meet growing demand for high-bandwidth memory and other AI-related memory chips. Such investment can help create production capacity, but the article gives no timeline for when new facilities or upgrades will increase output.

03

How did the companies’ combined cash and other liquid assets change over the past year?

Samsung and SK hynix together held 329.2 trillion won in cash and other liquid assets at the end of September. Their total had risen by 192.9 trillion won from 136.3 trillion won a year earlier. Liquid assets are cash or financial holdings that can be accessed relatively readily, so this buildup gives the companies substantial financial resources.

Samsung's cash equivalents and short-term financial products reached 214.8 trillion won, nearly double its 108.4 trillion won a year earlier. SK hynix's cash and other liquid assets more than quadrupled, rising from 27.85 trillion won to 114.5 trillion won. The combined total also grew by 51.32 trillion won during the third quarter alone.

The article says the buildup came alongside soaring profits from the AI memory boom, even as both companies made heavy investments. SK hynix also cut borrowings by 8.28 trillion won over the year, to 15.8 trillion won, and reported a net cash position of 98.66 trillion won. The figures describe their position at September's end, not a forecast.

04

What is high-bandwidth memory, and why are Samsung and SK hynix expanding production of it?

High-bandwidth memory, or HBM, is a form of memory built to transfer large amounts of data quickly. As general industry background, HBM uses stacked memory dies and a wide connection to processors; the article does not describe its construction. Its high data-transfer capacity makes it useful in systems that process demanding AI workloads.

AI processors need a steady flow of data to keep calculations moving. If memory cannot deliver data quickly enough, the processor may spend time waiting rather than doing useful work. The article identifies growing demand for HBM and other AI-related memory chips as the reason SK hynix is expanding production capacity. Both companies are investing in facilities and research as the AI memory boom lifts profits.

The article does not provide HBM production volumes or a measure of future demand. It does report that SK hynix invested 13.62 trillion won in the third quarter alone while expanding capacity to meet that demand. Samsung and SK hynix are both building up investment, but the figures do not show how quickly additional HBM supply will reach customers.

05

How could expanded chipmaking capacity affect the supply of memory chips for AI systems?

Expanded chipmaking capacity gives manufacturers the ability to produce more chips, which could help supply meet demand. That matters for AI systems because the article describes growing demand for high-bandwidth memory and other AI-related memory. However, investment in production facilities does not itself establish how many additional chips will be made.

SK hynix said it was expanding production capacity to meet growing demand, and it spent 13.62 trillion won on capital expenditure in the third quarter. Samsung also expanded and upgraded chip production facilities, with 52.75 trillion won of its nine-month capital spending going to its semiconductor division. These are concrete steps toward more or improved production capacity, though the article does not give expected output figures.

In general, added capacity can support greater supply once facilities and equipment are ready and production is running effectively. That is general industry context, not a timeline or result reported in the article. The current facts are that both companies are spending heavily and demand is growing; how much this will ease supply constraints, if any, is not specified.

06

Why does the AI boom require so much investment in both new factories and research and development?

The AI boom is increasing demand for memory chips, including high-bandwidth memory, according to the article. Meeting that demand calls for investment in production capacity and research. Factories and equipment support manufacturing; research and development helps companies create and improve the technologies they make. Together, these investments matter to companies seeking to serve a fast-growing market.

Samsung and SK hynix spent a combined 88.16 trillion won on capital expenditure in the first nine months. They also spent 51.89 trillion won on R&D. Samsung's R&D spending reached 41.8 trillion won, above its full-year 2025 total of 37.75 trillion won. SK hynix spent 10.09 trillion won, also above its full-year 2025 figure of 6.73 trillion won.

The article ties facility expansion to chip production and SK hynix's capacity increase to growing demand for HBM and other AI-related memory. It does not detail specific research projects or explain precisely which technical challenges the spending addresses. The current picture is record-level investment alongside rising cash reserves, fueled by soaring profits from the AI memory boom.

07

Why do AI processors need memory that can move data quickly, not just powerful computing chips?

A processor performs calculations, while memory stores data that the processor needs to use. In AI systems, keeping data flowing quickly can matter because a processor that waits for information cannot keep working at full pace. This basic explanation is general computing knowledge; the article does not describe processor-memory operation.

High-bandwidth memory is designed to move data at a high rate, which is why it is relevant to AI-related chips. The article reports growing demand for HBM and says SK hynix is expanding production capacity to meet it. It does not specify particular AI processors, workloads, or the technical limits that make fast memory necessary in a given system.

The article's focus is on the business response: Samsung and SK hynix invested heavily in chip facilities and R&D as the AI memory boom lifted profits. Their combined capital expenditure reached 88.16 trillion won in the first nine months, and combined R&D spending was 51.89 trillion won. These figures show the scale of investment, but do not quantify how much faster memory improves any particular AI system.

This brief was written by AI from the original reporting and checked by other models. Names, figures and quotes come from the source; read it for full context.

Read more in the JupiteX app

Pulse is free. New stories every 4 hours, each one broken into the questions that explain it.

Or read more news on the web