SK Square Posts Record Profit as SK hynix Powers AI Memory Gains

SK Square has returned to Google News with a record-profit headline, despite the underlying results dating to August 2025. That timing matters because the numbers describe an earlier stage of the AI memory boom, not a newly announced quarter.

The original report said SK Square generated 1.4011 trillion won in second-quarter operating profit. First-half operating profit reached 3.0534 trillion won, up 178% from the previous year. Strong earnings at SK hynix supplied much of that lift through equity-method accounting.

A year later, the relationship has become even more important. SK hynix has reported much larger profits, while Samsung has increased its advanced-memory production and China has expanded its semiconductor ambitions. The central question is no longer whether SK Square benefited from SK hynix. It is whether SK Square can turn that benefit into a broader, durable investment strategy.

The Google News Headline Comes From an Older Earnings Cycle

The record-profit headline is accurate within its original period, but it should not be mistaken for a new SK Square earnings announcement.

The underlying 2025 operating results were published on August 14, 2025. SK Square reported second-quarter operating profit of 1.4011 trillion won, an 80.8% increase from the same quarter a year earlier.

For the first half of 2025, the company recorded 809.4 billion won in sales. That represented a 16.2% year-over-year decline. Operating profit nevertheless increased 178%, from 1.0986 trillion won to 3.0534 trillion won.

That combination is the first clue to the story’s real mechanism. SK Square did not need rapid consolidated sales growth to produce a large profit increase. Its exposure to SK hynix allowed the memory manufacturer’s earnings to flow into SK Square’s results through equity-method gains.

Equity-method accounting records an investor’s share of an associated company’s profit. It does not mean the investment company received the same amount in cash during that quarter. This distinction matters when readers compare operating profit with money available for acquisitions, debt reduction, or shareholder distributions.

SK Square also credited tighter management of its information and communications technology portfolio. However, the scale and timing of the profit increase made SK hynix the central earnings driver.

The company was already returning capital to investors. It had started a 100 billion won share repurchase in April 2025 and said the repurchased shares would eventually be canceled. By then, cumulative repurchases and cancellations since SK Square’s November 2021 establishment had reached 610 billion won.

CEO Han Myung-jin connected those measures with a broader portfolio plan. He said SK Square was improving portfolio competitiveness, rebalancing assets, and preparing new investments focused on AI and semiconductors.

That strategy gave the results significance beyond one quarter. SK Square was positioning itself as both a major SK hynix shareholder and an investment vehicle for the next part of Korea’s technology buildout.

Google News can surface an older report beside newer coverage because aggregation systems respond to indexing, syndication, translation, and renewed reader interest. A headline’s appearance in a current feed does not change the event date inside the article.

Readers should therefore separate two timelines. The first is SK Square’s first-half 2025 performance. The second is the much larger semiconductor cycle that developed through 2026.

That separation does not make the older report irrelevant. It turns the report into a useful baseline for measuring how quickly SK Square’s dependence on SK hynix has increased.

SK hynix Turned an Investment Stake Into an Earnings Engine

SK Square’s profit surge reflects a structural exposure to SK hynix, not merely a successful quarter among several equal portfolio companies.

SK Square was created as an investment company after its separation from SK Telecom in 2021. Its portfolio included technology platforms, security businesses, digital commerce assets, and a major interest in SK hynix.

The semiconductor stake quickly became its defining asset. When SK hynix’s profit rises, SK Square recognizes its share through equity-method earnings. When the memory market contracts, the same accounting connection can work in reverse.

This arrangement explains why SK Square can report operating profit far above its consolidated revenue. The profit reflects income associated with its investment holdings, while revenue measures a narrower set of operating activities.

That structure became clear in SK Square’s full-year 2025 performance. The company later reported consolidated revenue of approximately 1.4 trillion won and operating profit of about 8.8 trillion won. Operating profit rose 124% from the previous year.

SK Square’s shareholder letter attributed the record performance to SK hynix, improved results across its ICT portfolio, and portfolio-rebalancing work. The letter presented those elements as parts of one investment strategy.

Still, the numbers establish a hierarchy. Portfolio improvements can remove losses and increase management flexibility. SK hynix can change the scale of SK Square’s entire income statement.

That dependence has advantages. SK Square offers investors indirect exposure to one of the leading suppliers of high-bandwidth memory, or HBM. HBM stacks multiple memory dies to feed data to AI accelerators at much higher bandwidth than conventional server memory.

HBM has become central to AI infrastructure because accelerators need to move model data rapidly between compute units and memory. A processor with inadequate memory bandwidth cannot use all its theoretical computing capacity efficiently.

SK hynix gained an early position in this market through close product development with major AI chip customers. That position helped the company benefit as cloud providers and model developers ordered more accelerators.

For SK Square, the result was an investment stake that behaved like an earnings engine. It also gave the company a stronger narrative for raising capital, recycling assets, and pursuing additional semiconductor investments.

The concentration carries a less flattering interpretation. SK Square’s record profit does not independently validate every business in its portfolio. It mainly demonstrates how valuable its SK hynix exposure became during an exceptional memory cycle.

That distinction affects how readers should evaluate future results. Improving a smaller portfolio company can create durable operating value. Recognizing another record profit from SK hynix can increase reported earnings without proving that SK Square has built a second comparable asset.

The company has tried to close that gap through portfolio rebalancing. It has sold or reorganized holdings that no longer fit its preferred direction and has emphasized AI and semiconductor investments.

Those actions are rational, but they also reveal the central challenge. SK Square must use the financial strength created by SK hynix without simply replicating its existing semiconductor exposure at higher valuations.

A successful investment company should convert one winning asset into several sources of future value. Otherwise, investors can bypass the holding company and buy the underlying semiconductor company directly.

AI Memory Demand Changed Korea’s Semiconductor Balance

The AI boom has shifted memory from a cyclical component business toward a strategic layer of computing infrastructure.

Traditional memory markets move through sharp supply cycles. Manufacturers expand capacity when prices and profits rise. New supply then meets slowing demand, pushing prices and margins down.

AI infrastructure has not eliminated that cycle. It has changed its shape by increasing demand for specialized products, advanced packaging, and long-term supply agreements.

SK hynix’s second-quarter results for 2026 show the scale of that change. The company reported revenue of 79.3187 trillion won and operating profit of 60.5426 trillion won.

Revenue rose 257% from the same quarter in 2025. Operating profit increased 557%, and the operating margin reached 76%. First-half revenue exceeded 100 trillion won for the first time in the company’s history.

SK hynix said high-value HBM, AI-server DRAM, and enterprise solid-state drives supported those results. Both DRAM and NAND prices also increased sharply from the previous quarter.

The company began mass shipments of HBM4 during the quarter and planned to increase production in the second half. HBM4 is a newer HBM generation designed to deliver greater bandwidth and efficiency for advanced AI systems.

SK hynix also said it had concluded long-term agreements with around ten customers. Such contracts can improve planning and reduce some exposure to short purchasing cycles, although they do not remove execution or pricing risks.

The company’s financial position strengthened alongside its profit. Cash and cash equivalents reached 88 trillion won at the end of the quarter, while total debt declined to 18.6 trillion won.

Those figures arrived after the older SK Square story surfaced through Google News. They show that the 2025 record was not the endpoint. It was an early marker in a much larger increase in AI-related memory earnings.

Samsung followed with its own extraordinary quarter. The company reported second-quarter 2026 operating profit of 89.5 trillion won and revenue of 171.5 trillion won.

According to an industry earnings review, Samsung and SK hynix together produce about two-thirds of the world’s memory chips. Their scale gives South Korea a central position in the AI hardware supply chain.

The competition between them also limits any simple victory narrative. Samsung has broad manufacturing capacity, a large customer base, and operations spanning memory, foundry services, processors, and consumer devices.

SK hynix has benefited from its position in advanced HBM. Samsung has responded by increasing HBM shipments and securing longer-term data-center supply arrangements.

Micron adds another major competitor. It has expanded its own HBM business and participates in the same qualification cycles for AI accelerators. Customers benefit from having multiple suppliers, particularly when memory shortages can delay entire server deployments.

Competition therefore occurs across more than headline market share. Memory companies must improve performance, power consumption, production yields, packaging capacity, and delivery reliability.

Yield measures the share of manufactured chips that meet required specifications. Poor yields can restrict supply and raise costs even when a company owns enough fabrication space.

Advanced packaging is equally important because HBM stacks memory dies and connects them to complex logic systems. A production bottleneck in packaging can constrain shipments even if wafer output increases.

This is why SK Square’s semiconductor push has strategic logic. AI infrastructure requires coordinated investment across memory, packaging, materials, data centers, and system design.

However, the same environment makes attractive investments expensive. Companies with credible HBM, packaging, or cooling technology already command high expectations. SK Square must distinguish genuine operating advantages from businesses merely using AI language.

Record Profit Does Not Remove the Concentration Risk

SK Square’s strongest asset is also the main source of uncertainty in its investment case.

An investment company normally reduces risk by holding several businesses whose results do not move together. SK Square’s recent performance has moved in the opposite direction because SK hynix has become more dominant.

The concentration is visible in reported earnings and market perception. Securities analysts have described SK Square as an alternative route to SK hynix exposure, especially when portfolio rules restrict direct allocations to one large stock.

A February 2026 valuation analysis noted that SK hynix had exceeded a 10% market-cap weight threshold. That development increased interest in SK Square as an indirect substitute.

The same analysis showed that SK Square continued to trade below the estimated net asset value of its holdings. A holding-company discount can reflect taxes, corporate structure, governance, capital-allocation risk, and uncertainty about when investors will realize the underlying value.

Share repurchases and cancellations can narrow that discount by reducing the share count. Asset sales can also help if management recycles capital into better opportunities or returns proceeds to shareholders.

Neither measure changes the underlying concentration by itself. If the value of SK hynix rises faster than every other asset, SK Square becomes more dependent on it even while the total portfolio grows.

There is also a gap between accounting profit and deployable cash. Equity-method profit increases reported earnings, but SK Square’s direct cash resources depend on dividends, financing, asset sales, and cash generated by consolidated businesses.

That gap matters when management promises new AI and semiconductor investments. Reported profit can make an investment company appear unusually well funded, yet the timing and size of available cash determine what it can actually buy.

The second risk comes from the memory cycle. AI spending has produced exceptional demand, but customers will eventually evaluate whether infrastructure investment generates acceptable returns.

A slowdown in accelerator orders would affect HBM suppliers. Additional capacity from SK hynix, Samsung, and Micron could also change pricing if production catches up with demand.

SK hynix says customer requests currently exceed supply capabilities. That statement supports continued investment, but it does not guarantee that every new facility will earn similar margins throughout its operating life.

Capital spending is rising across the industry. New fabrication plants and packaging facilities require long construction periods and large upfront commitments. They can begin production after the market conditions that justified them have changed.

Investor behavior already reflects this uncertainty. SK hynix and Samsung shares fell after their record second-quarter 2026 results because expectations had risen even faster than reported earnings.

That reaction is not evidence that the businesses are weak. It shows that record numbers can disappoint when the market has priced in even larger records.

Chinese competition creates another pressure point. Domestic memory producers have expanded capacity and technical capabilities, while Chinese authorities continue supporting semiconductor localization.

Export controls can slow access to some manufacturing equipment, but they can also motivate customers and governments to finance alternative supply chains. Over time, competition can emerge first in conventional memory before reaching more advanced products.

SK Square therefore faces two opposite risks. Moving too slowly could waste the financial advantage created by SK hynix. Moving too quickly could place capital into semiconductor assets near a cyclical peak.

The credible middle path requires disciplined investments with technical relevance, customer evidence, and a clear route to cash generation. Broad exposure to fashionable AI companies would not solve the concentration problem.

SK Square Must Prove Its Portfolio Can Produce a Second Winner

The next stage of the strategy depends on converting semiconductor gains into independent portfolio value.

SK Square has described AI and semiconductors as priorities for new investment. That direction fits South Korea’s industrial strengths and the demand created by global data-center construction.

The company does not need to find another memory manufacturer equal to SK hynix. It needs assets whose economics benefit from AI infrastructure without tracking the same risk in exactly the same way.

Advanced packaging is one possible area. HBM and AI accelerators require increasingly complex integration, testing, substrates, and thermal management. Companies that solve those constraints can benefit from rising system complexity.

Materials and manufacturing equipment offer another route. Memory production depends on specialized chemicals, deposition systems, inspection tools, and process-control software.

Data-center infrastructure also creates opportunities in power management, networking, storage, and cooling. These markets connect to AI demand but have different competitive structures and investment cycles.

SK Square’s challenge is to select where it possesses an information advantage. Its relationship with SK hynix can provide visibility into technical bottlenecks, supplier performance, and customer requirements.

That access is valuable only if governance protects both companies. SK Square must avoid transactions that transfer value between affiliates without clear benefits for each shareholder group.

Portfolio pruning is part of the same test. Selling a weak holding can improve future results even when the sale records a loss or attracts criticism.

The 2026 disposal of ONE store illustrated the tension. The Korean app marketplace had symbolic value as a domestic alternative to Google Play and Apple’s App Store, but it struggled to establish consistent profitability.

Employees criticized the proposed sale and questioned the buyer’s financial position. The episode showed that portfolio rebalancing can create labor, governance, and reputational costs.

Those costs do not automatically make a divestment wrong. They demonstrate why an investment company must explain how each sale supports its long-term capital plan.

SK Square should ultimately be judged on several linked outcomes. It must improve results at continuing portfolio companies, exit weaker positions responsibly, allocate cash at sensible valuations, and maintain shareholder discipline.

The company’s repurchase record supports the shareholder-return part of that case. Its record profit supports its access to valuable semiconductor exposure.

The unresolved part is repeatability. SK hynix’s rise reflects years of manufacturing investment, product qualification, and customer collaboration. A portfolio company cannot recreate that advantage simply by purchasing minority stakes in businesses labeled as AI suppliers.

New investments will need time before they produce material earnings. That delay can make the portfolio appear even more concentrated during the next several reporting periods.

Management should therefore disclose concrete progress where possible. Useful signals include realized gains, portfolio-company operating improvements, follow-on funding terms, customer contracts, and cash distributions.

Without those signals, investors must rely on announcements and estimated private-company valuations. Those measures are less informative than operating cash or independently priced transactions.

The strategic opportunity remains substantial. South Korea holds a major position in memory manufacturing and is increasing investments across semiconductor facilities and AI infrastructure.

SK Square sits close to that activity through its ownership, corporate relationships, and investment mandate. Few investment companies have comparable access to the operating center of the HBM market.

Access is not the same as execution. The next investment cycle must show that SK Square can identify constraints before they become consensus themes and support companies through commercialization.

Three Signals Will Decide What Happens Next

The next results, capacity decisions, and portfolio transactions will show whether this is a durable strategy or an amplified memory-cycle bet.

The first signal is SK Square’s next complete earnings disclosure. Readers should look beyond the operating-profit headline and separate SK hynix equity-method gains from improvements at other portfolio companies.

Growth outside SK hynix would strengthen the case that portfolio rebalancing is working. Another large profit increase without broader improvement would confirm that concentration remains the dominant story.

Cash movement deserves equal attention. Dividends received, proceeds from asset sales, repurchase spending, debt changes, and new investment commitments reveal more about capital-allocation capacity than accounting profit alone.

The second signal is SK hynix’s execution on HBM4 and new production capacity. The company has said HBM4 mass shipments started in the second quarter of 2026 and should increase during the second half.

Customers must qualify advanced memory before using it at scale. Stable yields, timely delivery, and repeat orders would support SK hynix’s claim that its HBM position rests on manufacturing execution rather than temporary scarcity.

Production plans will also test capital discipline. SK hynix is accelerating M15X production and preparing additional fabrication and packaging capacity.

Rising shipments without a sharp decline in pricing would strengthen the view that AI memory demand is structurally different from earlier cycles. Slower orders or aggressive price competition would weaken it.

The third signal is SK Square’s next material AI or semiconductor investment. The company needs a transaction that explains how SK hynix knowledge creates a genuine sourcing or evaluation advantage.

A credible deal would identify a specific infrastructure constraint, show customer validation, and establish how SK Square expects the investment to create value. A collection of small, loosely related AI stakes would offer weaker evidence.

The source of funding also matters. An investment financed through cash distributions or successful exits has different risk from one funded by rising leverage.

Google News readers encountering the resurfaced headline should keep these three signals in view. The original 2025 result captured the moment when SK hynix began transforming SK Square’s financial scale.

The larger question now concerns what management does with that advantage. SK Square can remain a discounted proxy for one valuable semiconductor company, or it can build a portfolio that earns its own strategic premium.

For developers and enterprise technology buyers, this matters beyond Korean equity markets. HBM supply affects accelerator availability, cloud capacity, infrastructure costs, and the timing of new AI deployments.

For knowledge workers tracking this market, the lesson is straightforward: preserve the event date, distinguish accounting profit from cash, and connect each headline with subsequent disclosures. A current Google News result can be useful, but only when its original timeline remains visible.

Watch the next SK Square filing, SK hynix’s HBM4 ramp, and the investment company’s next semiconductor deal. Together, those events will show whether record profit became a foundation for diversification or merely a larger exposure to the same cycle.