On April 29, 2026, Samsung Electronics (005930.KS) released a quarterly financial report that could redefine the history of Korean corporate performance.
Consolidated revenue reached KRW 133.87 trillion (approximately USD 90.1 billion), up 69.2% year-over-year and 42.7% quarter-over-quarter, setting a new record for single-quarter revenue. Operating profit soared to KRW 57.23 trillion (about USD 38.4 billion), skyrocketing 756.1% year-over-year and 185.1% quarter-over-quarter, easily beating both internal company forecasts and the market consensus of KRW 55.28 trillion.
To put this in perspective: Samsung’s total operating profit for all of 2025 was KRW 43.6 trillion. This means that just one quarter in 2026 has already surpassed the entire previous year’s total.
Within the profit structure, the Device Solutions (DS) division, responsible for semiconductor operations, emerged as the absolute profit engine. The DS division posted KRW 53.7 trillion in operating profit for the quarter, accounting for roughly 93.9% of the company’s total operating profit—up from just KRW 1.1 trillion in the same period last year. With an operating margin of about 65.7%, DS was the driving force behind Samsung’s overall operating margin leap to roughly 42.8%.
Following the earnings release, several international investment banks promptly raised their target prices. Goldman Sachs increased its target price for 005930 from KRW 205,000 to KRW 260,000, maintaining a "Buy" rating. Citi, on May 12, raised its target price from KRW 300,000 to KRW 460,000 and initiated a 90-day upward catalyst watch. Nomura went even further, lifting its target price from KRW 340,000 to KRW 590,000 on May 18.
However, what truly prompted market reflection wasn’t the fully priced financial figures, but a casual remark from management during the earnings call: At this point in time, traditional DRAM is actually more profitable than HBM.
This statement directly contradicts prevailing market assumptions. In the narrative of the "AI memory supercycle," HBM is seen as the most profitable and highest-barrier segment. Why would management signal such a "counterintuitive" message during a record-breaking quarter? The underlying business logic is far more complex than the financial data alone.
A "Counterintuitive" Profit Structure
The assertion that traditional DRAM is more profitable than HBM isn’t an isolated stance by Samsung—it’s supported by clear industry data.
According to the latest report from semiconductor research firm TrendForce, published on May 27, 2026, annual pricing mechanisms and other factors led to HBM wafer value and profit margin being overtaken by DDR5 RDIMM in Q1 2026. TrendForce had already noted in January that server DRAM unit prices and margins had surpassed HBM, making it the most profitable item.
Specifically, HBM’s pricing cycle differs structurally from standardized DRAM. HBM contracts are typically negotiated annually, with longer lock-in periods and price adjustments lagging behind the spot market. Traditional DRAM (such as DDR5), on the other hand, uses a more responsive price discovery mechanism, quickly reflecting supply-demand gaps and price increases.
Price data makes this contrast even clearer. In Q1 2026, general DRAM contract prices jumped 90% to 95% quarter-over-quarter, while NAND Flash prices rose 55% to 60%. With HBM’s large-scale ramp-up occupying much of the available capacity, traditional DRAM supply shrank even faster than demand grew, creating a positive spiral: Scarcity drives up prices, and higher prices boost profits.
The global DRAM market’s performance confirms this trend. According to Counterpoint Research data released on May 27, 2026, global DRAM revenue in Q1 surged 80% quarter-over-quarter and 260% year-over-year, reaching a record USD 97 billion—just shy of the USD 100 billion mark in a single quarter. Samsung maintained its lead with a 38% share of global DRAM revenue.
Behind this inverted profit margin lies a deeper capacity logic: the capacity squeeze effect. HBM die sizes are significantly larger than those of comparable standard DRAM, consuming much more wafer capacity. As HBM shipments ramp up, advanced process capacity is heavily occupied, causing effective supply of traditional DRAM to shrink. The combination of supply-demand imbalance and flexible pricing mechanisms ultimately allows traditional DRAM to surpass the relatively rigidly priced HBM in marginal profitability.
The HBM Segment: Competitive Realities Beneath the Halo
To understand the profit inversion between traditional DRAM and HBM, we need to examine the competitive landscape of the HBM market.
For the critical HBM4 generation, Samsung achieved the world’s first mass production and commercial shipment in February 2026. The company expects HBM sales in 2026 to more than triple last year’s total, with HBM4 revenue accounting for over half of all HBM sales starting in Q3.
On the technology front, Samsung accelerated development of HBM4E. At the GTC conference in March 2026, Samsung publicly showcased HBM4E physical samples, featuring a per-pin speed of 16 Gbps and total bandwidth exceeding 4.0 TB/s—about 21% higher than HBM4. The first batch of HBM4E engineering samples began production in Q2, prioritized for major customers conducting system-level validation.
Yet market share data paints a different picture. TrendForce estimates SK Hynix will hold about 50% of the HBM market in 2026, with Micron increasing to 28%. Counterpoint Research’s global DRAM revenue rankings show Samsung leading with 38%, followed by SK Hynix at 29% and Micron at 22%.
Although Samsung enjoys a timing advantage with the first mass production of HBM4, SK Hynix maintains a lead in supply allocation thanks to its longstanding HBM partnership with NVIDIA. This "first-mover barrier" isn’t solely determined by technical specs, but by lengthy certification cycles, joint validation history, and accumulated supply chain trust. Whether Samsung’s catch-up strategy can achieve a breakthrough in HBM4 and subsequent generations depends on its ability to convert mass production leadership into substantial customer supply share while keeping pace technologically.
The Pricing Power Game: Profit Lag Under Annual Agreements
TrendForce’s May 2026 report highlights the core contradiction in HBM pricing: HBM contracts for 2026 use an annual pricing framework, with prices locked in as early as the second half of 2025. When the overall DRAM market entered a supply crunch in Q1 2026 due to surging AI demand, HBM contract prices remained "pinned" at relatively low levels, unable to rise in tandem with standardized DRAM.
Additionally, 2026 saw new suppliers enter the HBM market in earnest, increasing supply and giving buyers more leverage in contract negotiations, further limiting short-term HBM price upside.
This lagging pricing framework has triggered strong reactions from suppliers. TrendForce notes that as negotiations for 2027 HBM4 supply begin in Q2 2026, manufacturers will substantially raise HBM prices to reflect supply-demand imbalances and new-generation manufacturing costs. The rising profitability of traditional DRAM gives suppliers greater bargaining power—if HBM profits remain lower than traditional products, manufacturers lack economic incentive to expand HBM supply, forcing buyers to accept higher HBM contract prices.
SK Hynix’s strategy diverges here. According to TrendForce, with RDIMM prices overtaking HBM, SK Hynix stated it "does not pursue short-term profits, but seeks balanced long-term development between HBM and conventional DRAM." This differentiated stance suggests Samsung and SK Hynix may be diverging in capacity allocation strategies: Samsung appears more inclined to maximize short-term gains from high-profit traditional DRAM, while SK Hynix focuses on consolidating HBM’s long-term supply position.
May 2026 Labor Strike: Supply Risks in the Supercycle
In an already tight capacity environment, Samsung Electronics faced the largest labor conflict in its history in May 2026. The National Samsung Electronics Union announced an 18-day general strike plan, demanding 15% of annual operating profit be allocated to employee performance bonuses, while management only agreed to 10% and refused to make the bonus system permanent.
On May 18, the Suwon District Court partially granted Samsung’s injunction request, ordering that safety-related personnel and duties be maintained during the strike, and production materials not be allowed to deteriorate. Any violations would incur a daily fine of KRW 300 million. Labor-management mediation broke down on May 20, but both sides eventually reached a temporary wage agreement later in May. Following a union-wide vote, 73.7% of members approved the deal, averting the planned large-scale strike.
The root cause of the strike lies in the imbalance of benefit distribution during the AI supercycle. Samsung’s operating profit surged 756% in Q1, but chip division employees felt they hadn’t received a commensurate share of the AI-driven profit boom. Meanwhile, SK Hynix reached an agreement with its union in September 2025, abolishing the previous bonus cap (formerly 1,000% of base salary) and allocating 10% of annual operating profit to the bonus pool for all employees, effective for ten years. Estimates suggest SK Hynix employees could receive KRW 600 million to KRW 700 million in bonuses in 2026, while Samsung chip division employees in equivalent roles receive less than a third of that. This pay gap has led about 200 key Samsung engineers to move to SK Hynix in recent months.
From an industry perspective, any subsequent strike that disrupts production would quickly ripple through the global semiconductor supply chain. JPMorgan estimates that a prolonged strike at Samsung could impact 2026 operating profit by KRW 2.1 trillion to KRW 3.5 trillion, and drag semiconductor division revenue by 1% to 2%. KB Securities in Korea projects that if only 30% to 40% of union members participate, it could disrupt about 3% to 4% of global DRAM supply. With the supply-demand gap widening, any real capacity loss would further drive up memory chip prices.
The Three Layers of the Supercycle
Synthesizing all these dimensions, we can clearly identify the three-layer structure of the current AI memory supercycle:
Layer One: Structural Demand Drivers. Unlike previous short-term recoveries driven by consumer electronics upgrade cycles, this memory market rally is fueled by structural demand reassessment from AI training, inference, and large-scale data center expansion. Nomura Securities notes in its research that as AI semiconductor demand shifts from training to inference workloads, memory demand is entering a period of exponential growth. Inference scenarios’ KV-cache memory demand is a multiplicative function of user base, engagement time, AI task complexity, and token consumption—meaning memory demand could grow thousands-fold over the next five years.
Layer Two: Capacity Squeeze and Supply Rigidity. Large-scale HBM expansion is occupying vast amounts of advanced wafer capacity, directly squeezing traditional DRAM supply. After the previous downturn, memory manufacturers have become more cautious in capital expenditure, and new capacity release cycles typically take 12 to 18 months. IDC China’s VP Zhou Zhengang predicts this memory price upcycle will span 2026 and could extend into 2027 or 2028.
Layer Three: Redistribution of Pricing Power. As cloud service providers sign long-term supply agreements with locked-in bottom prices, overall DRAM pricing in 2026 exhibits a "hard to rise, harder to fall" structure. Meanwhile, HBM’s annual pricing framework is facing fundamental reset, with 2027 contract negotiations expected to deliver price hikes far beyond previous market expectations. TrendForce explicitly states that manufacturers will dynamically balance capacity allocation based on profitability comparisons between HBM and traditional DRAM.
The interplay among these three layers defines the unique nature of this supercycle: It’s not just a demand-driven boom, but a "disciplined prosperity" constrained by supply rigidity—manufacturers, aiming to maximize profits, are incentivized to maintain rather than disrupt the current supply-demand structure. This makes the supercycle’s duration and stability sharply different from the traditional "boom-surplus-crash" memory cycles of the past.
Conclusion: The "Internal Tension" of the Supercycle
Samsung’s Q1 2026 financial report appears, on the surface, to be a perfect illustration of the "AI memory supercycle"—record-high revenue and profit, doubling growth rates, and HBM shipments expanding threefold. Yet beneath these dazzling numbers, management’s comment that "traditional DRAM is currently more profitable than HBM" reveals an overlooked internal tension within the supercycle narrative.
This tension manifests on multiple fronts: the inverted profit structure between traditional DRAM and HBM, the lagging annual pricing mechanism of HBM and the imminent contract price reset, the ongoing rivalry between Samsung and SK Hynix in market share and technology cadence, and the tug-of-war between AI-driven profit surges and internal benefit distribution. These tensions are not negations of the supercycle; on the contrary, they are key clues to understanding its unique characteristics.
From an investment perspective, the current valuation narrative for 005930 is at a pivotal transition point. Once the explosive profit growth is fully priced in, the next focus will shift to the health and sustainability of the profit structure—whether HBM contract prices can catch up in 2027, whether labor risks can be institutionally resolved, and whether HBM supply share can achieve a breakthrough. The evolution of these variables will determine whether Samsung Electronics can transition from a "cycle beneficiary" to a "structural leader" in the AI memory supercycle.
