News · Data & AI Infrastructure
Samsung's 19-Fold Profit Jump Signals Rising AI Memory Costs for Enterprises
Samsung expects second-quarter operating profit of about 89.4 trillion won, 19 times last year's level, as AI data centers drive up memory prices. Enterprise buyers should expect those costs to show up in server and cloud bills.
By The AI Enterprise Magazine Team4 min readNews · Weekly briefing
Key takeaways
- Record guidance. Samsung forecast about 89.4 trillion won in Q2 operating profit, up from 4.7 trillion won a year earlier and above consensus.
- Memory prices are surging. Citi data cited by The Next Web showed DRAM contract prices up about 44% and NAND up about 53% quarter on quarter.
- Investors are wary. Samsung and SK Hynix shares fell sharply as analysts questioned how long hyperscaler spending can last.
- Budget impact. Enterprises should expect higher memory costs to flow into server quotes, GPU cloud pricing and storage renewals.
Week of July 6, 2026
Samsung's Quarterly Operating Profit Soars on AI Memory Demand
Samsung's preliminary second-quarter operating profit reached about 89.4 trillion won, up from 4.7 trillion won a year earlier, as AI data centers drove memory prices higher.
View the data
| Operating profit | |
|---|---|
| Q2 2025 | ₩4.7T |
| Q1 2026 | ₩57.2T |
| Q2 2026 | ₩89.4T |
Samsung Electronics said on Tuesday that it expects operating profit of about 89.4 trillion won, roughly $58.4 billion, for the April to June quarter, about 19 times what it earned a year earlier. The driver is the price of memory chips that AI data centers cannot get enough of. For enterprise technology leaders, the number is less a story about Samsung than a signal about what AI infrastructure will cost for the rest of 2026.
The preliminary figure, disclosed in a filing with South Korean regulators, came in 6.2% above market forecasts, Il Sole 24 Ore reported, citing Yonhap. A year ago, Samsung earned 4.7 trillion won in the same quarter, and in the first three months of 2026 it earned 57.2 trillion won, according to Analytics India Magazine.
What did Samsung report for the second quarter?
Revenue is expected to reach about 171 trillion won, more than double the 74.57 trillion won of the prior-year quarter, The Next Web reported. That was slightly below the 173 trillion won some analysts had modeled, but operating profit beat a consensus of about 84 trillion won. It is Samsung's third record quarter in a row.
The headline number may understate the chip business. Samsung's guidance reportedly includes a large provision for employee bonuses, and without it operating profit would have topped 100 trillion won, according to analysts cited by Reuters and the Seoul Economic Daily. Samsung has not yet disclosed net profit or a breakdown by division. Full results are due July 30.
Analytics India Magazine set the figure next to Nvidia's most recent quarterly operating income of about $53.5 billion, a comparison that would make Samsung one of the most profitable technology companies in the world for the quarter. A cleaner comparison will have to wait for Samsung's net profit.
Why are memory prices rising so fast?
AI servers need far more memory than ordinary servers, and suppliers have not kept up. Contract prices for DRAM rose about 44% from the previous quarter and NAND flash prices rose about 53%, according to Citi Research data cited by The Next Web. The pressure started with high-bandwidth memory (HBM), the stacked chips placed beside Nvidia's accelerators, but it has spread to conventional DRAM and NAND as manufacturers moved capacity toward HBM and as AI workloads pushed into storage and general server buildouts.
Samsung is not alone. Micron's revenue roughly quadrupled in its June report, The Next Web noted, and SK Hynix was scheduled to begin trading American depositary shares on the Nasdaq on July 10 in a listing of roughly $28 billion, according to Analytics India Magazine.
Why did Samsung's shares fall on a record quarter?
Investors did not celebrate. Samsung shares fell as much as 7.9% in early trading, SK Hynix fell as much as 7.3% and the KOSPI index dropped nearly 6%, Analytics India Magazine reported. Analysts raised doubts about whether hyperscale cloud providers can keep spending on AI infrastructure at the current pace, and whether the returns will justify the industry's capital plans.
That tension matters to enterprise buyers. If the cloud providers keep building, memory stays tight and expensive. If they slow down, prices could ease, but so could the pace at which new GPU capacity reaches the market.
What does this mean for enterprise AI budgets?
Higher component prices rarely stay with the chipmaker. Enterprises should expect the effects to show up in on-premises server quotes, in the pricing of GPU cloud instances and in storage contracts that renew in the second half of the year.
Three practical steps follow. First, lock in pricing on hardware orders that are already approved, since contract prices have been resetting each quarter. Second, revisit capacity plans for private AI infrastructure, where memory and storage costs are now harder to forecast. Third, treat inference efficiency as a cost control, not just an engineering goal, because every unnecessary token and oversized model consumes memory somewhere.
Speakers at AI Connect events have warned that AI infrastructure costs grow quietly. Without monitoring, auditing and alerts, "in six months you might see a bill" as large as several years of past spending, Varun Tottempudi, head of data and platforms engineering for global analytics at Lockton, said at COLLIDE 2024, in a session on AI ROI across the project life cycle. For teams weighing their own hardware, the HPE and NVIDIA view of the AI factory and a AI Connect panel on controlling GPU and inference costs offer a useful checklist.
What should leaders watch next?
The next markers are Samsung's full results on July 30, which will show how much of the profit came from memory, and the big cloud providers' next updates on capital spending in the coming weeks. Those reports will show whether the buildout that is driving memory prices is still accelerating, and how long enterprise buyers should plan for tight supply.
Frequently asked questions
How much profit did Samsung forecast for Q2 2026?
About 89.4 trillion won, roughly $58.4 billion, in preliminary operating profit, about 19 times the 4.7 trillion won it earned a year earlier.
Why are memory chip prices rising?
AI servers need large amounts of memory, and suppliers shifted capacity to high-bandwidth memory, tightening supply of conventional DRAM and NAND as well.
How does this affect enterprise AI costs?
Higher memory prices tend to pass through to server hardware, GPU cloud instances and storage contracts, so enterprises should lock in approved orders and revisit capacity plans.
Sources
Filed under: Data & AI Infrastructure, AI Strategy & ROI
