Samsung in the Shadow of a Record Profit: Why the Market Stays Wary at the Peak of AI Memory

Samsung Electronics has guided that its third-quarter (July–September) operating profit will reach 107.4 trillion won — roughly 80.17 billion US dollars. That is close to nine times the figure from the same quarter a year earlier, and a rare number in Korean corporate history. Yet in the same period the company's share price has fallen more than 25 percent below its June record high. On the KOSPI index, this decline and the guided record profit are happening at once, and that divergence is the most important signal of the moment. The LSEG SmartEstimate put the analyst consensus at 106.1 trillion won, meaning Samsung has beaten expectations only modestly. That is the first signal: for the market, this result is not a surprise but the closing of an expected chapter. Revenue is large too — around 195 trillion won for the quarter, roughly 127 percent growth year on year. This is Samsung's fourth consecutive quarter of record operating profit. Looking backwards, the run is impressive. But the question now points forward, not backward. Background: a record resting on one division At the centre of the record sits memory. Samsung itself says the bulk of the earnings improvement came from its memory division. Demand from artificial-intelligence infrastructure — data centres and AI accelerators in particular — has pushed up prices for high-bandwidth memory (HBM), DRAM and NAND flash. That price surge has temporarily concentrated pricing power in the hands of a few firms: Samsung, SK Hynix and Micron. Analysts describe it as a kind of oligopoly profit, where tight supply leaves buyers with few alternatives. But the profit has an uncomfortable shape. Nearly all of Samsung's earnings now depend on a single division. While memory breaks records, the mobile division lost more than a billion dollars in the third quarter, and the foundry business — making chips for others — remains loss-making because factory utilisation is low. The headline number looks vast, but its foundation is not broad. A car running on one engine stops when the fuel runs out; if memory prices fall one day, much of this profit could vanish. Those two loss-making divisions are not merely an expense line; they raise a question about the quality of the profit. When investors see a record number, they also ask whether it is durable or merely a snapshot taken at the top of the cycle. For Samsung the answer is ambiguous. Memory is earning, but other parts of the company are eating that gain from within. The HBM race: how close has Samsung come Samsung's biggest strategic story is now HBM, the scarcest and most expensive component in AI accelerators. SK Hynix has long led this market, and Samsung is chasing. According to the company, HBM bit shipments rose about 50 percent quarter on quarter, showing Samsung is not merely a bystander — it can narrow the gap to the leader. But narrowing a gap is not the same as closing it. SK Hynix leads not only on technology; it has built long-term relationships and trust with leading AI chipmakers. Those links take time to break or replace. The real test for Samsung is therefore not just raising output but establishing itself as a reliable supplier to major customers. A 50 percent quarterly jump is encouraging, but without consistency it is only one good quarter. Here Samsung's strategic calculus gets complicated. It is strong in conventional DRAM and NAND, but future profit depends on the high value-add of HBM. If it reaches SK Hynix sooner than expected, the base of its earnings broadens. If it stays behind, today's record profit will remain a temporary image of an uncertain future. Price growth is decelerating The most important signal hidden inside the numbers is the change in momentum. According to TrendForce market research, DRAM contract-price growth was around 60 percent in the second quarter; for the fourth quarter it is forecast to fall to 10–15 percent. Prices are still rising, but not as fast. Similarly, fourth-quarter profit growth may slow to 8.2 percent quarter on quarter, far below the 20 percent of the third quarter. That deceleration means the cycle is probably approaching a peak. Memory markets have always moved in waves. AI-driven demand is strong now, and analysts say the supply-demand imbalance could persist into 2028. If that forecast holds, Samsung stays advantaged. But a forecast is never a certainty. If supply rises quickly, the shortage could resolve earlier, and margins would come under pressure. Two questions now face Samsung. First, how much profit survives if memory prices start to fall. Second, how quickly that profit can absorb the losses in mobile and foundry. The answers together will shape Samsung's next year. Two loss-making divisions and the arithmetic of money Samsung's problem lies not in the size of the profit but in its composition. The mobile division's quarterly loss has exceeded a billion dollars, partly because rising memory prices raise the cost of making the company's own devices. This is the most curious situation: Samsung earns by selling memory, and also buys that same memory to build devices, so its own costs rise from within. Foundry is similar. Low utilisation means fixed costs cannot be spread, so the division stays in the red. In this market TSMC is the clear leader and Samsung the trailing challenger. Demand for advanced-node chips could turn the situation around, but no one knows how fast that recovery will come. From the outside, the two divisions are now leaning on memory profits. Anyone judging Samsung by the total number alone would miss the real picture, because one record figure can hide both a strong engine and two weak wheels. The competitive map Samsung's position in semiconductors is now spread across several fronts. In HBM it is advancing fast; in conventional DRAM and NAND it is broadly at parity; but in foundry it has a structural gap to TSMC. At the same time, Chinese memory makers are growing stronger, adding supply at lower prices, which could pressure the pricing power of Samsung, SK Hynix and Micron over the longer term. This competition means today's record profit is the product of a special situation: tight supply and intense demand. Over time supply will rise, rivals will multiply, and pricing power will weaken. Samsung's strategic challenge is therefore not merely producing more, but securing its place in the high-value segment like HBM, where competition is thinner and margins more durable. This raises the question of how quickly Chinese rivals can enter complex technologies like HBM. If they do so quickly, Samsung's window shrinks. If slowly, Samsung can capture the full benefit of a shortage that may last to 2028. Much of the long-term earnings outlook hinges on this single variable. What is happening across the value chain The memory wave is not confined to chipmakers; it spreads across the entire technology chain. Prices rise upstream, and the shock travels downstream, raising costs for smartphone, PC and consumer-electronics makers. Samsung stands at both ends, so profit and loss grind against each other inside the same company. AI infrastructure demand is the biggest driver of this chain. The expansion of data centres, cloud services and AI model training has lifted memory demand. But how long that demand lasts at this pace depends on the investment plans of large technology firms. If AI investment slows, the first blow lands on memory prices, and the first earnings impact hits suppliers like Samsung. For this reason Samsung's profit is not solely the fruit of its own skill but part of a global technology cycle. Chip prices, AI investment and currency rates — these three external variables now drive the company's income. The larger the external variables, the greater the uncertainty. Another quiet risk in the arithmetic A further quiet pressure comes from currency markets. Samsung's overseas sales are in dollars but reported in won. When the won strengthens against the dollar, the won value of the same dollar sales falls, trimming reported income and prompting analysts to cut forecasts. This currency risk is not dramatic, but it slowly erodes the earnings picture. Notably, Samsung has limited natural protection, and an export-dependent company cannot fully shield itself when currencies swing. A contrarian thought: is the story looking backwards There is a trap in how this record profit is being told. Headlines speak of a world first, vast profit, an AI boom. But that language is built by looking backwards, at what has already happened. The market is speaking a very different language. A share price down more than 25 percent from its high means investors are pricing forward risk, not backward results. Here lies the biggest lesson. A record number never guarantees the future by itself. Samsung has posted record operating profit for four straight quarters — robust evidence in the backward view, but a single up-cycle cannot prove forward durability. Four good quarters mean four good quarters, nothing more. Another point stands out — the gap between media heat and market behaviour. Celebration in the headlines, scepticism on the exchange. This divergence often appears at the top of a cycle. When everyone agrees a boom is underway, the boom is frequently in its final phase. In Samsung's case the market has already given that cautionary signal, even as headlines still sing victory. Analyst Kim Seok-hwan of Mirae Asset Securities notes that the market's real question is no longer how big the profit is, but how long it will last. Anyone offering certainty on that answer is probably underestimating the market. Looking ahead: October 29 and beyond The next key moment in Samsung's story is the detailed earnings release on October 29, when not only the total but each division's figures will appear. If mobile and foundry losses are larger than expected, that could add further pressure on the share price. If the company clarifies shareholder-return policy — a buyback or dividend — market sentiment could turn. Three paths are open to Samsung at once. One, as long as memory prices stay high, profit stays strong. Two, Chinese competition and the pace of AI investment could end that advantage sooner than expected. Three, progress toward HBM leadership could keep Samsung standing beyond the cycle. Which path becomes real depends on numbers no one can yet state with certainty. A profit figure is never permanent. What Samsung has achieved is the fruit of a particular moment — intense demand, tight supply, high prices. When that moment passes, the question will remain: was the company merely standing at the crest of a wave, or has it built a foundation to stand after the wave? October 29 may offer the first clue. The complete answer will come only with time, exactly when the next cycle's pricing arithmetic begins.