By NUPIAO
The intense demand for DRAM and NAND memory is pushing Samsung Electronics toward what could be its most jaw-dropping profit forecast ever.
On July 6, according to LSEG’s SmartEstimate model, which aggregates forecasts from 30 analysts, Samsung’s Q2 2026 operating profit could hit 86 trillion won (about $563.5 billion), an 18-fold jump from 4.7 trillion won a year earlier. That would mark the third straight quarter of record-breaking profit. Samsung Electronics is set to release its preliminary Q2 earnings on Tuesday.
Right now, the global memory chip market is still in a structural supply shortage. Analysts at several firms believe that the expansion of AI inference infrastructure is outpacing the production capacity of major memory vendors like Samsung, SK Hynix, and Micron, and this supply-demand imbalance could last at least through 2027.
Unlike earlier cycles driven only by HBM, this memory price rally is lifting all boats. The rise of agentic AI is broadening computing demand, servers need larger memory to handle multi-step complex tasks, and the massive data storage needs of inference are boosting demand for traditional DRAM and NAND flash. Samsung, as a core storage supplier to Nvidia, Google, and Apple, is catching all that incremental demand across the value chain.
Citi Research’s price monitoring report shows that average DRAM prices rose 44% quarter-on-quarter in Q2, while NAND flash prices jumped 53%. These across-the-board price hikes are directly padding Samsung’s quarterly profit margin.
So far this year, Samsung Electronics, SK Hynix, and Micron have seen their stock prices surge 158%, 273%, and 242% respectively. All three memory titans have surpassed a market cap of $1 trillion, making the memory sector the hottest asset class in global tech.
Nomura Securities is even more bullish in its latest quarterly report, forecasting that commodity DRAM prices will rise another 24% sequentially from July to September, and NAND flash will climb 25%. The triple demand from consumer electronics, traditional data centers, and AI computing is expected to keep chip prices elevated.
Samsung is fast becoming the go-to foundry for Big Tech’s custom AI chips (ASICs), with a long-term order backlog approaching 50 trillion won. On July 3, news broke that Meta is pushing forward with Samsung’s foundry to co-design and produce next-gen ASICs worth over 10 trillion won. Its in-house AI accelerator “MTIA” has already locked in Samsung as a partner, planning to mass-produce hundreds of thousands of units using the advanced 2nm process. Meanwhile, U.S. AI giant Anthropic is also evaluating Samsung’s 2nm technology for its own chip development.
On top of that, Samsung’s foundry unit has recently tweaked its supply strategy, prioritizing existing customers and selectively taking new orders. Its 4nm process capacity is pretty much sold out—even for next year—and some 8nm lines are running near full tilt.
Facing the long-term demand for AI memory, on June 29, Samsung and SK Hynix jointly announced a massive 3,200 trillion won (about $2.07 trillion) domestic chip expansion investment plan. Samsung’s investment spans from 2026 to 2040, focusing on expanding its Pyeongtaek and Yongin wafer clusters, adding a new advanced manufacturing base in Gwangju, and ramping up HBM advanced packaging, next-gen DRAM, and NAND capacity. SK Hynix didn’t share a detailed timeline but said it would speed up the construction of its Yongin memory base, moving up the completion of a fab originally planned for 2045 by twelve years.
Such a massive long-term investment also means hefty capital spending for years, which will keep siphoning off operating cash flow. And there’s another thing that could weigh on Samsung’s profit: a huge bonus provision.
Back in May, Samsung reached a wage deal with its semiconductor division union, defusing the biggest strike threat in the company’s 50-plus-year history. The agreement sets aside 10.5% of the semiconductor unit’s operating profit as a special bonus pool for chip employees, with no cap on payouts.
Several brokerages estimate that if the full provision is recognized in Q2, Samsung’s cumulative bonus reserve could exceed 40 trillion won. The timing of this mega-expense will directly shrink the reported operating profit, and in an extreme case, the actual results might miss the market’s consensus forecast of 86 trillion won.
Meanwhile, a J.P. Morgan research note points out that the share of AI storage in cloud providers’ capital spending is climbing fast—expected to hit 52% this year and break 70% next year. The big debate is whether that high share can be sustained. If AI companies pull back on hardware spending, memory demand could cool quickly, and Samsung and SK Hynix’s massive expansion plans could face overcapacity risks.
Investors are now waiting for real proof that AI services are commercializing and that all this computing investment will translate into cloud and AI revenue growth. That’s what’s needed to justify the high valuations of memory chips. Until concrete data arrives, the sector faces potential valuation pullbacks.
It’s worth noting that Samsung’s mobile division is also feeling the cost squeeze. Higher memory chip prices are pushing up the overall component costs for smartphones, wiping out the revenue gains from recent price hikes. As a result, the mobile unit’s profit margins are shrinking.
Analysts expect Samsung to raise smartphone prices again later this year to offset the cost pressure. Its biggest rival, Apple, already hiked prices across the iPad and MacBook lineup last month. While the wave of consumer electronics price hikes is industry-wide, continued increases could hurt product competitiveness—another challenge Samsung will have to navigate.
