Record-breaking profits but hidden risks: Samsung Electronics AI storage cycle's three major uncertainties

In the first quarter of 2026, Samsung Electronics turned in a performance worthy of the history books. Both revenue and profit set new record highs, HBM4 reached mass production as scheduled, and demand for AI memory seems to have no ceiling.

However, when you look through the footnotes to the financial report and the records of the earnings call, a more complex picture emerges—while the entire market is cheering for HBM, Samsung’s management disclosed a counterintuitive fact: the profitability of traditional DRAM today is actually higher than that of the celebrated HBM. During the call, Samsung explicitly stated, “The profit margin of ordinary DRAM is actually higher than HBM.” The reason is a difference in pricing mechanisms: ordinary DRAM prices are adjusted through quarterly negotiations, allowing them to quickly reflect changes in supply and demand, whereas HBM prices are locked in via annual contracts.

At the same time, although a strike crisis involving more than 75,000 contract workers was temporarily resolved at the last minute through an interim agreement and thus appeared to be put on hold, the structural disputes between labor and management over profit distribution have not truly been settled. This casts a continuing layer of uncertainty over the critical window for capacity ramp-up and customer validation.

This is not just Samsung’s story—it is a snapshot of the entire AI hardware infrastructure entering deep water.

A Financial Report That Breaks Every Expected Upper Limit

On April 30, Samsung Electronics released its official financial report for the first quarter of 2026, with data that comprehensively exceeded market expectations.

Consolidated revenue reached 133.9 trillion won, up 69% year over year and 43% quarter over quarter. Even more striking to the market was operating profit—57.2 trillion won, up as much as 756% year over year and 185% quarter over quarter. Net profit also climbed to 47.2 trillion won. All three core indicators set new single-quarter records since Samsung Electronics was founded in 1969. Even operating profit alone in the first quarter surpassed the full-year 2025 level by 43.5 trillion won.

By segment, the Device Solutions segment—i.e., the semiconductor business—contributed 93.9% of total operating profit for the quarter. In the quarter, the DS division posted revenue of 81.7 trillion won and operating profit of 53.7 trillion won, with an operating margin of about 66%. AI-related DRAM and HBM products are the core driving forces.

However, not every segment enjoyed the growth dividend. The Device Experience segment recorded consolidated revenue of 52.7 trillion won and operating profit of 3 trillion won. While the mobile experience business benefited from stable sales of the Galaxy S26 Ultra flagship, rising component costs constrained profit growth.

The core engine behind this earnings report is the fifth-generation high-bandwidth memory HBM4, which began mass production and shipments in February. According to the disclosure, HBM4 capacity is already fully booked for all of 2026, and Samsung expects 2026 HBM sales to exceed 2025 by more than three times. Samsung also confirmed that it will send HBM4E engineering samples to core customers in the second quarter.

In terms of institutional ratings, KB Securities raised Samsung’s target price sharply to 530,000 won on May 28. Citibank increased its target price from 300,000 won to 460,000 won on May 12, about 60% higher than the then-closing price. As of the end of April, 37 analysts covering Samsung Electronics had an average target price of 274,603 won.

Data Cut: The Surface and the Hidden Layers of Profit Structure

The fact that the DS segment accounts for 93.9% of Samsung’s profit contribution clearly shows how heavily Samsung’s current performance depends on the semiconductor business.

The key information disclosed by Samsung during the earnings call—“traditional DRAM is more profitable than HBM”—reveals deeper characteristics of the profit structure. This runs counter to market intuition: despite HBM having higher technical barriers, its current gross margin is lower than that of traditional DRAM.

The reason lies in a fundamental difference in the pricing mechanism. Traditional DRAM prices are negotiated quarterly, enabling them to reflect supply-and-demand shifts quickly; meanwhile, HBM prices are locked in through annual contracts. In the current cycle where storage prices are continuously rising, quarterly-priced traditional DRAM can benefit from price increases faster. A surge in data center demand has led to tight DRAM supply. With HBM taking up most of the capacity, supply of other DRAM types is further squeezed, pushing the average selling price of general-purpose products up significantly.

For investors, this means two things. First, HBM’s long-term pricing power remains solid, but the pace of short-term profit realization may be slower than expected. Second, once traditional DRAM prices enter a downcycle, the magnitude of fluctuations in Samsung’s profits could exceed market expectations.

Profit compression in the mobile business points to another dimension. Revenue of 52.7 trillion won in the DX division contributed only 3 trillion won in operating profit. As the world’s largest smartphone manufacturer, Samsung’s cost structure is facing challenges. Rising component costs and tariff burdens are the main sources of pressure.

Viewpoint Clash: Divergence Under the “Supercycle” Consensus

The market is not unified regarding Samsung’s current valuation and future trajectory.

The bullish case has clear and convincing arguments. Full-year booking for HBM4 throughout 2026 means the lower bound for 2026 semiconductor business revenue has effectively been locked in. The delivery of HBM4E samples soon after in the second quarter also provides visibility into growth for 2027. KB Securities raised its target price to 530,000 won and Citibank raised its target price to 460,000 won—both well above the average of analysts. Some institutions forecast that Samsung’s full-year 2026 operating profit will reach 327 trillion won, and that in 2027 it is more likely to rise to 488 trillion won.

More aggressive views come from some buy-side institutions. They argue that key competitors are still optimizing yields at the HBM3E stage, giving Samsung an important window during the HBM4 generation. Samsung regained the leading position in overall DRAM market share by the end of 2025. Since HBM4 began mass production in February, performance has exceeded early expectations, and full-year 2026 capacity has already been completely sold out.

Cautious voices are also worth paying attention to. Their concerns concentrate on three areas. First, the drivers of the current DRAM cycle are highly concentrated in a single demand source—AI servers. If cloud providers slow their capital expenditure pace, demand gaps could widen quickly. Second, because HBM follows an annual pricing model, Samsung may be unable to fully capture the flexibility of spot price increases, and the spot prices of traditional DRAM already show signs of having peaked. Third, although the strike crisis has been temporarily resolved, structural tensions between labor and management still create long-term uncertainty. Industry observers note that if a full-scale strike lasts 18 days, potential losses could be as high as 30 trillion won. JPMorgan, from the perspective of additional labor costs, also estimates that in the worst case, Samsung’s 2026 operating profit could face a 12% downside risk.

Ripple Effects: From South Korea’s Semiconductors to the Global AI Hardware Chain

The impact of Samsung’s earnings report extends far beyond the single-company level.

From the perspective of the competitive landscape, differences in the mass production progress of HBM4 are reshaping the global HBM market structure. Samsung’s lead in entering the HBM4 generation for mass production, while main competitors are still optimizing yields at the HBM3E stage, means that the supply map of HBM for AI accelerators such as NVIDIA’s next-generation GPU architectures could shift significantly compared with the previous generation.

From the perspective of industry chain transmission, Samsung’s full-year HBM capacity booking sends a clear signal: advanced packaging capacity is likely to remain tight in the second half of 2026. Packaging—the intermediary layer between HBM and GPUs—is one of the bottlenecks for AI chip capacity. The alignment between Samsung’s HBM shipment schedule and packaging capacity will directly affect the delivery timeline of AI servers worldwide.

On a more macro level, Samsung’s Q1 performance reinforces the market’s judgment that there is a positive AI capital expenditure cycle. When upstream memory suppliers record record profit growth, it implies that cloud providers are paying higher prices to secure supply. These costs ultimately need to be absorbed by revenue growth at the AI application layer. If monetization at the application level does not keep up with the growth rate of infrastructure investment, the cycle could face a stress test at some point.

For the crypto and Web3 space, the industry trend also has a corresponding logic. The crossover narrative between AI and crypto—including decentralized compute markets, AI Agent token economics, and on-chain inference verification—depends on the continued expansion of underlying AI hardware and cost reductions. Samsung’s HBM4 mass production progress and pricing trends will indirectly influence the cost curves and adoption pace in these tracks.

Conclusion

Samsung Electronics’ performance in the first quarter of 2026 nearly fulfills all the market’s expectations for an “AI memory supercycle.” Revenue, profit, and profit margins simultaneously set record highs; HBM4 mass production is underway with full-year booking; and the product roadmap for the next generation is clearly visible. However, what is truly worth thinking about in this earnings report is not just the height of the numbers, but the structural signals behind them.

The fact that traditional DRAM profit margins surpass HBM reveals an asymmetric effect of the pricing mechanism during cycle upswings. When the market generally fixates on the most cutting-edge HBM, it is often the “traditional” product lines that determine short-term profit elasticity. This finding reminds investors that technological leadership does not necessarily translate into profit leadership. In the strong-cycle memory chip industry, getting the timing right can sometimes matter more than choosing the direction.

The temporary resolution of the strike crisis does not eliminate the underlying tension between labor and management regarding profit distribution. During high-profit cycles, employees’ demands to share in the growth dividend are naturally reasonable, and balancing this with maximizing shareholder returns will become a structural issue Samsung faces over the medium and long term.

Samsung’s Q1 performance is undoubtedly a concentrated reflection of demand for hardware infrastructure in the AI era. But the highly concentrated nature of that demand is also a side of cyclical vulnerability. When a company’s fate is deeply tied to the capital expenditure timing of a small number of key customers, the room for valuation fluctuation will inevitably expand. Understanding this is more helpful for investors than simply remembering the figure of 57.2 trillion won, as it can help them navigate through the fog of the cycle.