Samsung Electronics is poised to deliver another historic quarter, underscoring how artificial intelligence has transformed the global semiconductor industry from a cyclical business into one driven by relentless demand for computing power.
The South Korean technology giant is expected to report an operating profit of 86 trillion won ($56.35 billion) for the April to June quarter, almost 18 times higher than the 4.7 trillion won recorded a year earlier, according to an LSEG SmartEstimate based on forecasts from 30 analysts. If realized, the result would mark Samsung’s third consecutive quarter of record operating profit, extending one of the strongest earnings streaks in the company’s history.
The expected performance comes as AI investment continues to stretch global memory supplies, allowing manufacturers to command significantly higher prices for memory chips that have become essential components of modern AI infrastructure.
For Samsung, the boom represents a remarkable turnaround. Just a few years ago, the world’s largest memory chipmaker was battling one of the industry’s worst downturns as oversupply and weak electronics demand crushed chip prices. Today, the opposite problem confronts the industry. Demand has overtaken supply, inventories have tightened, and memory prices have surged as cloud providers and technology companies race to build AI infrastructure.
Analysts believe the favorable supply-demand imbalance is unlikely to disappear anytime soon, with expectations that memory shortages will persist through at least next year.
The AI revolution has changed not only the volume of memory being consumed but also the type of demand driving the market. Early waves of generative AI were primarily centered on training massive language models, creating enormous demand for high-bandwidth memory (HBM), the premium memory technology used alongside AI accelerators. That trend continues, but analysts say a broader shift is now underway.
The rapid expansion of inference computing, where trained AI models respond to user requests in real time, is creating strong demand for conventional DRAM and NAND memory products that form the backbone of servers and storage systems.
The emergence of agentic AI is accelerating that trend.
Unlike traditional AI chatbots that simply generate responses, agentic AI systems execute complex, multi-step tasks, retrieve information from multiple sources, make autonomous decisions and continuously interact with software applications. Those capabilities require larger memory pools for server processors and significantly greater storage capacity to retain, retrieve, and process vast amounts of data efficiently.
As enterprises deploy sophisticated AI agents across business operations, demand is expanding beyond specialized AI chips into virtually every category of memory semiconductor.
That places Samsung in an enviable position.
The company supplies memory chips to many of the world’s biggest technology companies, including Nvidia, Google and Apple, making it one of the biggest beneficiaries of the global AI infrastructure race.
The pricing environment exposes just how tight supplies have become. According to Citi Research, average selling prices for DRAM jumped 44% quarter-on-quarter during the April to June period, while NAND flash prices rose an even steeper 53%.
Such increases would have been almost unimaginable during previous semiconductor cycles, when price swings were typically driven by fluctuations in smartphone or PC demand. Today, AI infrastructure spending has become the dominant force shaping memory markets.
The Electronics Side Too
The earnings boom has also fueled one of the biggest stock market rallies in the technology sector.
Samsung Electronics shares have surged approximately 158% this year. Rival SK Hynix has climbed about 273%, while U.S. memory producer Micron Technology has gained roughly 242%. The extraordinary rally has pushed the market valuations of all three memory manufacturers above the $1 trillion mark, highlighting investors’ confidence that AI-driven demand will remain robust for years rather than quarters.
Yet Samsung’s headline earnings could still contain one important caveat. Analysts caution that reported operating profit may come in below market expectations if the company recognizes larger employee bonus provisions during the quarter.
In late May, Samsung reached a wage agreement with its semiconductor workers, ending the threat of a large-scale strike that had raised concerns about disruptions to production. Under the agreement, 10.5% of the semiconductor division’s operating profit will be allocated as special bonuses for chip employees.
Some analysts estimate cumulative bonus provisions could exceed 40 trillion won. While those payments would not alter the underlying strength of Samsung’s semiconductor business, the timing of recognizing the expense could materially influence reported quarterly earnings.
Samsung is expected to publish its full earnings report later this month, offering investors greater clarity on how those accounting provisions affect the final figures.
But even as the industry enjoys record profitability, analysts are increasingly focused on one question that could determine how long the boom lasts: Can AI investment continue growing at its current pace?
JPMorgan recently said investor discussions have shifted away from whether memory demand remains strong and toward whether the industry’s biggest customers can continue allocating such a large share of their capital spending to AI infrastructure. The bank estimates AI memory already accounts for approximately 52% of cloud service providers’ capital expenditure this year and expects that figure to exceed 70% next year.
That concentration has prompted growing debate over whether spending can remain sustainable without corresponding growth in commercial AI services. Investors are now seeking evidence that AI products are generating enough revenue to justify continued investment in massive data centers, specialized processors and increasingly expensive memory systems.
Those concerns carry particular significance because memory manufacturers are embarking on some of the largest investment programmes in the industry’s history.
Last week, Samsung and SK Hynix announced plans to invest a combined 3,200 trillion won ($2.07 trillion) to expand semiconductor production capacity in South Korea. Samsung expects to spread its investment between 2026 and 2040, while SK Hynix has not disclosed a detailed implementation schedule.
The scale of those commitments reflects confidence that AI demand will remain structurally higher for many years. However, it also raises the financial stakes should cloud providers slow capital spending or delay planned AI deployments.
To reduce uncertainty, Samsung has already begun securing future demand through long-term customer agreements. The company disclosed in April that it had signed multi-year binding contracts with customers seeking guaranteed memory supplies, although it did not reveal either the identities of those customers or the financial terms of the agreements.
The outlook for pricing remains favorable.
Nomura expects commodity DRAM prices to rise another 24% quarter-on-quarter during the July to September period, while NAND prices are projected to increase 25%, supported by stronger demand from consumer electronics manufacturers as well as operators of conventional and AI-focused data centers.
The expectation of another quarter of price increases suggests the industry’s supply constraints remain far from resolved.
Gain Here, Pain There
Ironically, Samsung’s success in semiconductors is creating new challenges elsewhere within the company. Its smartphone business is increasingly feeling the impact of soaring memory costs, with more expensive components squeezing margins even after recent handset price increases.
Analysts say higher semiconductor costs have more than offset Samsung’s pricing adjustments, reducing profitability in the mobile division.
That pressure could intensify during the second half of the year if memory prices continue climbing.
Some analysts believe Samsung may need another round of smartphone price increases to protect margins, mirroring moves already taken by Apple, which raised prices on selected iPad and MacBook models last month.
The divergence between Samsung’s businesses illustrates how profoundly AI has reshaped the technology industry. The semiconductor division is generating unprecedented profits because memory has become one of the most valuable components of AI infrastructure. At the same time, those same elevated memory prices are increasing manufacturing costs for smartphones, tablets, and other consumer devices.
