Our Team's Analysis | Samsung Electronics Q2 Earnings Review: After Record Profits, the Market Questions Sustainability

1. Record-breaking results: Revenue slightly below expectations, profit in line with high-cycle assumptions

$Samsung Electronics (005930.KR)$ The Q2 earnings guidance itself was very strong. The company expects consolidated revenue of approximately KRW 171 trillion and operating profit of about KRW 89.4 trillion for the quarter. According to sell-side estimates, this implies year-over-year revenue growth of roughly 129.3% and sequential growth of about 28%, while operating profit surged 1,810.2% year-over-year and rose 56% sequentially. The sharp year-over-year increase in operating profit reached a historic high, continuing the earnings recovery trend driven by the upward phase of the memory cycle over recent quarters. Samsung will release its full financial report on July 30, which will include detailed segment breakdowns, gross margins, and management guidance.

However, from the buyer’s perspective, this guidance was not 'broadly ahead of expectations.' Market reaction was muted, primarily becauserevenue came in slightly below some market expectations, while profitability—though strong—was largely in line with investor assumptions about a super-cycle in memory markets. According to certain consensus estimates, Samsung’s Q2 revenue was expected to be around KRW 173–174 trillion, whereas the company guided to KRW 171 trillion, slightly below consensus. For operating profit, the market expected approximately KRW 87 trillion, and the company guided to KRW 89.4 trillion, modestly above expectations. In other words, the earnings composition was as follows:Revenue fell short of full expectations, profit came in slightly better than expected, and margins remained elevated.

It should be noted that different institutions use varying definitions for consensus estimates: UBS, in its July 3 preview, cited a consensus estimate of roughly KRW 169 trillion in revenue and KRW 87 trillion in operating profit, while publicly reported figures referenced estimates closer to KRW 173.9 trillion in revenue and KRW 87.3 trillion in operating profit.

This reflects two underlying signals.

First, memory pricing and product mix are indeed very strong. Demand from AI servers has driven sustained price increases in DRAM, HBM, server memory, and enterprise SSDs. As one of the world’s largest memory manufacturers, Samsung directly benefits from tight industry supply-demand dynamics. In particular, DRAM and NAND exhibited significant price elasticity in Q2, substantially boosting memory segment margins. Citi estimates Samsung’s Q2 semiconductor operating profit at approximately KRW 88.4 trillion, display at KRW 0.5 trillion, mobile at KRW 0.3 trillion, consumer electronics at KRW 0.06 trillion, and other segments at KRW 0.15 trillion—indicating nearly all profit was contributed by the semiconductor division.

Second, strong margin performance does not imply all business segments beat expectations. Slightly lower-than-expected revenue suggests the volume-price combination did not fully surpass the market’s most optimistic assumptions; meanwhile, operating profit modestly exceeding expectations indicates favorable pricing, cost structure, and product mix. Importantly, profit was also affected by bonus provisions. Korean media previously reported that Samsung might have accrued several trillion won in bonus reserves due to labor agreements and performance-based compensation arrangements; international outlets also noted that employee bonus accruals partially offset Q2 profits. Citi estimates Q2 employee bonus provisions at approximately KRW 17.2 trillion, while Morgan Stanley estimates bonus provisions accounted for roughly 10% of operating profit and may include one-time recognition factors related to the first half of the year.

Therefore, if the impact of bonus provisions is excluded, underlying operating profit for Q2 could be even stronger. However, capital markets did not simply interpret this as an 'above-expectations positive,' because the market had already fully priced in DRAM price increases, HBM volume ramp-up, and AI server demand pull-through. In other words, Q2 results confirmed a strong cyclical upswing but failed to deliver sufficient new information to justify further earnings upgrades. This is also the core reason behind Samsung's share price decline following its earnings release: it wasn't poor performance, but rather the market’s shift in focus from 'record-breaking profits' to 'whether further upside revisions are still possible.'

II. What the Market Is Worried About: Price Momentum, HBM Delivery, and AI Capital Expenditure

After Samsung's Q2 results, buy-side investors care less about the quarter’s profit and more about three forward-looking variables.

The first variable is whether memory prices can continue rising in Q3. Q2 was likely the fastest phase of the current memory price rally. Previously, TrendForce projected that Q2 2026 contract prices for mainstream DRAM would rise 58%–63% quarter-over-quarter, while NAND Flash contract prices would increase 70%–75%. These gains were extreme, driven by supply constraints on standard DRAM and NAND caused by capacity reallocation toward HBM, server DRAM, and enterprise SSDs. However, the issue is that even if prices continue rising in Q3, the pace of increase will likely slow significantly. Subsequent guidance from TrendForce suggests Q3 DRAM contract prices may rise only 13%–18%, and NAND Flash contract prices 10%–15%. In short, prices are still climbing, but the slope has shifted from 'explosive growth' back to 'moderate increases.' Record-high contract prices are now approaching the affordability limits of PC and smartphone customers, which is a key reason behind the notable slowdown in Q3 price momentum.

This point is critical for memory stocks.Stocks trade not on absolute price levels, but on quarter-over-quarter changes in average selling prices (ASPs) and the pace of earnings upgrades. When DRAM and NAND prices surged sharply in Q2, the market continuously revised profit forecasts upward; but if price increases start moderating in Q3—even while prices remain elevated—the market will worry that the momentum for further earnings upgrades is fading. Thus, the biggest near-term divergence on Samsung isn’t whether memory prices will drop imminently, but whether the steepest phase of price increases is already behind us. The more reasonable buy-side view is that prices may not reverse, but the rate of increase has likely peaked.

The second variable is whether HBM supply progress can continue to exceed expectations. Samsung’s recent valuation recovery stems not only from rising traditional DRAM and NAND prices but also from market expectations around its HBM catch-up. HBM is one of the highest-margin and strategically most valuable memory products in AI servers, and it represents the core battleground where Samsung competes with SK Hynix and Micron. The July 30 earnings call highlighted three key points worth tracking: the impact of bonus provisions, memory supply-demand dynamics from H2 2026 into 2027, and further progress on long-term agreements (LTAs). Meanwhile, delays in key customer certifications for HBM represent a primary downside risk. If Samsung’s HBM supply execution proceeds smoothly—particularly if it secures additional certifications and orders from major clients for high-end HBM products—the market would assign it higher earnings and valuation elasticity. The rationale is straightforward: traditional DRAM price hikes generate cyclical profits, but HBM volume ramp-up signals an improved structural position for Samsung in AI memory. Conversely, if HBM supply progress fails to further surpass expectations, the market may conclude that Samsung is primarily benefiting from legacy memory price increases rather than gaining meaningful share in high-end AI memory.

The third variable is whether the AI-related capital expenditure boom will marginally cool off in the coming quarters. Current strength in memory demand is fundamentally driven by AI infrastructure build-out. Expansion by North American cloud providers, AI model companies, and the GPU and ASIC supply chains has boosted demand for HBM, server DRAM, high-capacity RDIMMs, and enterprise SSDs. As long as AI-related capex continues to rise, demand for Samsung’s high-end memory products remains supported. However, the market is now starting to worry about another issue: whether AI capex has already entered an overheated phase. Over the past few quarters, investment in AI data centers has grown very rapidly, with expansions across storage, GPUs, networking equipment, power infrastructure, and data center facilities. If cloud vendors begin prioritizing return on investment, depreciation pressures, and compute utilization in the coming quarters, the marginal growth rate of AI-related capex could slow. AI memory is expected to account for 52% of cloud vendors’ capex this year and could exceed 70% next year; what the market truly worries about is not short-term demand, but whether such a high allocation can be sustainably validated (JPMorgan).This does not necessarily imply a collapse in demand, but it is already significant enough to affect memory stock valuations. Given the inherently cyclical nature of the memory industry, once the market perceives a slowdown in demand growth—while supply is gradually released between 2027 and 2028—it will start pricing in risks of supply-demand rebalancing or even oversupply ahead of time.

Therefore, the post-earnings share price adjustment for Samsung reflects not operational issues in the current quarter, but the market’s forward-looking discussion of three medium-term questions: whether prices can keep rising in Q3, whether HBM can meet heightened expectations, and whether AI-related capex will marginally cool off.

III. Summary: Fundamentals remain strong, but extrapolation of price increase momentum should be tempered

From an investor perspective, Samsung should not be simplistically viewed as bearish at this point, nor should the earnings elasticity seen in Q2 be linearly extrapolated. A neutral assessment is:Samsung’s fundamentals remain in a strong cycle, but its share price has now entered a phase of earnings realization and sustainability validation. On the optimistic side, memory prices have not yet reversed course. DRAM and NAND prices are still rising, and server DRAM, HBM, and enterprise SSDs remain the tightest segments in terms of supply-demand balance. Industry capacity is being absorbed by high-end AI products, squeezing supply for mainstream DRAM and NAND, which likely continues to support Samsung’s Q3 earnings. On July 3, TrendForce cited supply chain sources indicating that Samsung is negotiating with customers to raise average DRAM selling prices in Q3 by up to 20%, with LPDDR potentially seeing increases exceeding 20%. Bonus provisions partially offset Q2 profits, further indicating that underlying operational profitability remains robust.

On the cautious side, Q2 was likely the quarter with the strongest sequential ASP (average selling price) elasticity. Even if prices continue to rise going forward, the magnitude of increases could notably diminish. For equities, a deceleration in price gains—from 60%–70% down to 10%–20%—does not signal a fundamental reversal, but it is sufficient to dampen market expectations. What the market truly fears is this scenario: profits remain high, but the pace of upward revisions slows; prices keep climbing, but the rate of increase peaks; demand stays strong, yet AI-related capex starts facing scrutiny. Meanwhile, industry media reports show that Samsung Electronics, SK Hynix, and Micron have seen their share prices rise approximately 158%, 273%, and 242% year-to-date, respectively—indicating that markets have already fully priced in the previous optimistic narrative of 'AI plus memory shortages.'

Therefore, an investment framework of 'strong fundamentals, cautious on chasing rallies' is more appropriate.

First, in the short term, post-earnings declines should not be interpreted as a fundamental reversal. Samsung's Q2 results still indicate a very strong memory cycle, with AI server demand continuing to materialize and margins remaining elevated. As long as DRAM and NAND prices keep rising in Q3 and server orders show no significant weakening, the stock price could find support after the recent correction.

Second, medium-term assumptions about continuously surging average selling prices (ASP) need to be tempered. The magnitude of price increases seen in Q2 is unlikely to be sustained over the long term. Going forward, the focus should shift to whether price gains are gradually moderating quarter-over-quarter, rather than solely on whether prices are still rising. If DRAM and NAND price increases continue to decelerate beyond Q3, the market may reframe Samsung from a 'continuously upwardly revised earnings trade' to a 'sustained high-margin trade,' reducing valuation elasticity.

Third, whether Samsung can regain upside potential hinges on HBM. Traditional memory price increases generate cyclical profits, whereas HBM customer certifications and rising supply share drive structural re-rating. If Samsung continues to deliver on HBM—particularly by narrowing the gap with SK Hynix in high-end customers and next-generation products—the market could once again revise its medium- to long-term earnings trajectory upward.

Fourth, it’s essential to monitor marginal changes in AI-related capital expenditures. If North American cloud providers continue to raise their AI data center investments, demand for server DRAM, HBM, and enterprise SSDs could remain robust for longer. However, if capital spending slows, AI compute leasing prices weaken, or data center ROI comes under scrutiny in the coming quarters, memory stocks would likely face valuation compression first.

Earnings remain strong and fundamentals have not reversed, but the easiest phase—driven by expanding expectations—may already be behind us.

The views expressed herein are solely my own. Investing involves risks; please bear all risks yourself.

30K Views