Samsung’s Q2 Operating Profit Soared 1,810% on AI Memory Chips

Samsung Electronics expects to book a record 89.4 trillion won ($58.4 billion) in second-quarter operating profit, up 1,810% from a year earlier, as AI-driven demand for high-bandwidth memory lifted memory chip prices and volumes at the same time. The preliminary guidance, released Tuesday, also puts revenue at 171 trillion won, up 129.3% from a year earlier, and marks Samsung’s third consecutive quarter of record results.

The scale of the beat, $58.4 billion in three months on a single product category, comes with an asterisk investors noticed immediately. The same companies whose memory chip allocations have powered the AI build-out are now trimming output of the chips that go into phones, laptops and consumer electronics, and prices for those devices are already moving up.

Samsung Posts Its Biggest Quarter Ever on AI Memory Demand

• 89.4 trillion won Q2 operating profit estimate ($58.4 billion)
• Up 1,810% from 4.68 trillion won a year earlier
• 171 trillion won in revenue, up 129.3% on year
• 6.2% above the 84.16 trillion won Yonhap Infomax consensus
• Third consecutive quarter of record results

The $58.4 billion figure is Samsung’s largest quarterly operating profit to date, beating the 84.16 trillion won average estimate compiled by Yonhap Infomax by 6.2% and Bloomberg’s 84.2 trillion won estimate by a similar margin. The Korea Herald, reporting the preliminary filing on Tuesday, called the jump the largest on record, and Yonhap confirmed the same numbers in its own summary of the regulatory filing.

A year earlier, Samsung’s Q2 operating profit was 4.68 trillion won, a small base that magnifies the 19-fold gain. The 89.4 trillion won estimate includes provisions for the special chip-division bonuses Samsung agreed to in May; excluding those, the underlying figure would reach around 100 trillion won, Yonhap reported. Under the wage deal, Samsung grants a special semiconductor performance bonus equal to 10.5% of business performance earnings, paid in company stock over at least 10 years and tied to performance targets in the chip business.

It is Samsung’s third consecutive quarterly record, following the fourth quarter of 2025 and the first quarter of 2026, and the final audited numbers are due July 30. Revenue, while more than doubling, came in just shy of the 173.3 trillion won consensus, a detail the stock noticed.

Why Memory Prices Are Climbing So Fast

Memory prices have climbed because AI data centers are pulling in chips at a rate the industry’s existing fabs cannot match. Hyperscalers, including Microsoft, Google, Meta and Amazon, are pouring capital into AI accelerators, and each accelerator stacks large quantities of high-bandwidth memory on top of Nvidia GPUs. That demand pulled the wider memory market into a supply squeeze that has now lasted several quarters and shows no sign of breaking.

Samsung has positioned itself as the lead supplier of the newest generation. In February, it became the first company in the industry to mass-produce and ship sixth-generation HBM4 chips, built on a sixth-generation 10-nanometer-class 1c DRAM process with a 4-nanometer logic die. Samsung’s HBM4 mass-production announcement said the company took the unconventional route of using its most advanced node on the memory side to hit performance targets from day one of mass production.

HBM4 delivers 11.7 gigabits-per-second, exceeding the 8 Gbps industry standard by roughly 46%, and Samsung expects its HBM sales to more than triple in 2026 versus 2025. Industry observers quoted by Yonhap expect favorable memory conditions to continue at least through next year.

Instead of taking the conventional path of utilizing existing proven designs, Samsung took the leap and adopted the most advanced nodes like the 1c DRAM and 4nm logic process for HBM4.

Sang Joon Hwang, Executive Vice President and Head of Memory Development at Samsung Electronics, made the comment in the company’s HBM4 announcement.

The Hidden Loser in the AI Memory Boom

The global memory shortage analysis and reallocation thesis from IDC put the trade-off plainly. Major memory makers have shifted production toward memory used in AI data centers, such as high-bandwidth memory and high-capacity DDR5, the firm wrote, and this has restricted the supply of general-purpose memory modules and driven up prices across the board. The result, IDC argued, is a permanent reallocation of the world’s silicon wafer capacity rather than a temporary mismatch.

The shift is already biting Samsung’s own consumer division. The company’s device experience unit, which runs its mobile and consumer electronics business, is expected to post relatively weak results in Q2, and its foundry business is expected to clock potential losses from the higher bonus expenses it agreed to in May. Samsung also makes consumer phones and laptops, so the same supply squeeze that powered its chip profit is now a headwind for its own devices.

This is a zero-sum game: every wafer allocated to an HBM stack for an Nvidia GPU is a wafer denied to the LPDDR5X module of a mid-range smartphone or the SSD of a consumer laptop.

IDC analyst team, in a global memory shortage analysis.

How Much More Phones and PCs Will Cost in 2026

Memory is one of the most expensive components in a finished phone. IDC estimates it at 15-20% of the total bill of materials for a mid-range device and 10-15% for a flagship. When memory prices rise across the board, that line item pulls the whole bill up with it.

PC vendors are already passing the cost through. Lenovo, Dell, HP, Acer and ASUS have warned clients of tougher conditions ahead, confirming 15-20% hikes and contract resets as an industry-wide response, IDC reported. The pressure comes at a particularly bad time for the PC market, where the memory shortage is colliding with the Microsoft Windows 10 end-of-life refresh cycle and the broader AI PC marketing push.

IDC framed two downside scenarios for 2026, one moderate and one pessimistic, depending on how long the supply tightness lasts:

Vendor exposure is uneven across the device industry:

• Most exposed: thin-margin vendors TCL, Transsion, Realme, Xiaomi, Oppo, Vivo, Honor and Huawei, where memory cost rises will likely be passed straight through to end users.
• Already raising prices: Lenovo, Dell, HP, Acer and ASUS, with 15-20% hikes and contract resets confirmed.
• Hedged: Apple and Samsung, protected by cash reserves and long-term supply agreements that lock memory 12 to 24 months in advance.

Korea’s 800 Trillion Won Chip Bet and the Tax Windfall

South Korea is staking a national bet on the cycle continuing. On June 29, President Lee Jae Myung announced a combined 800 trillion won chip hub with Samsung and SK hynix each set to build two new semiconductor fabs in the country’s southwest. Samsung is allocating 400 trillion won for new fabs in Gwangju and 56 trillion won for HBM facilities in Cheonan and Onyang. SK hynix is putting 400 trillion won into a southwest production base and pulling forward completion of its fourth Yongin fab to 2033, twelve years earlier than previously planned.

The two companies together make about two-thirds of the world’s memory chips, according to the AP, so the cluster is designed to sit at the center of the AI hardware supply chain rather than at its margins. The HBM angle is central: AI accelerators consume stacked memory in large volumes, and HBM production uses wafer capacity that would otherwise serve more conventional DRAM markets.

The boom is also producing a tax windfall. Analyst estimates cited by Yonhap suggest Samsung and SK hynix will together generate more than 100 trillion won, or about $65 billion, in annual corporate taxes. Presidential chief of staff Kang Hoon-sik said Sunday that the government plans to channel the excess into a Future Response Fund that would back three national mega projects, housing, startups and jobs for people in their 20s and 30s. A broader Seoul package, announced the prior week, anticipates at least 1,350 trillion won, roughly $880 billion, of investment from companies including Samsung and SK hynix.

Why Investors Sold Into a Record Print

The print was the largest quarterly operating profit Samsung has ever reported, but the shares moved the other way. Samsung stock slid as much as 8% in morning trade on Tuesday, dragging the KOSPI index down more than 6%, and rival SK hynix tumbled 7%.

Three reasons drove the sell-off:

• Revenue came in just shy of the 173.3 trillion won consensus, the opposite of a clean beat on the top line for a quarter that had set up as pristine.
• Profit-taking: Samsung shares have more than doubled in 2026, and the guidance triggered the kind of move that follows a long run.
• Sustainability doubts: investors are weighing how long AI-driven memory demand will last, even as industry observers said favorable conditions are likely to continue through next year.

Samsung will release its final Q2 numbers on July 30, when the audited figures resolve which of these readings holds.

Frequently Asked Questions

Why did Samsung’s Q2 2026 operating profit jump 1,810%?

AI-driven demand for high-bandwidth memory lifted both volumes and prices at the same time. Samsung estimated 89.4 trillion won in operating profit on 171 trillion won in revenue, with the device solutions unit, which houses its semiconductor business, accounting for the dominant share of the gain.

How much more will phones and laptops cost in 2026?

IDC’s moderate scenarios have smartphone average selling prices up 3% to 5% and PC average selling prices up 4% to 6%; its pessimistic scenarios have smartphone prices up 6% to 8% and PC prices up 6% to 8%. Lenovo, Dell, HP, Acer and ASUS have already confirmed 15-20% price hikes and contract resets.

Is the AI memory boom sustainable?

Industry observers quoted by Yonhap expect favorable memory conditions to continue at least through next year. Beyond that horizon, the cycle is no longer a temporary mismatch but a strategic reallocation of wafer capacity to AI customers, according to IDC, which means consumer electronics will keep paying higher memory prices until new fabs come online.

What is the 800 trillion won chip hub?

Announced June 29 by President Lee Jae Myung, it is a combined 800 trillion won plan under which Samsung and SK hynix will each build two new semiconductor fabs in South Korea’s southwest, adding capacity at the center of the AI hardware supply chain. Samsung plans 400 trillion won for new fabs in Gwangju and 56 trillion won for HBM facilities in Cheonan and Onyang.

Why did Samsung shares fall on a record profit print?

Revenue came in just shy of the 173.3 trillion won consensus, Samsung shares have more than doubled so far in 2026 and were vulnerable to profit-taking, and investors are weighing whether AI-driven memory demand can hold beyond next year.