On October 28, 2026, SK hynix announced that its third-quarter revenue reached 104.6155 trillion won, the first time quarterly revenue has topped 100 trillion won. Operating profit also hit 83.3338 trillion won, and both figures set new records. The operating margin reached 80%.
Backed by these record results, the company is expanding long-term supply contracts and boosting production capacity. The number of long-term supply customers has grown from about 10 in July to 14, and the company says it has secured contract volumes equal to more than 50% of its total production capacity for 2027 and 2028. Securing supply volume, however, is a separate matter from fixing sales prices. The contracts include a mechanism that reflects market conditions within a price range agreed with each customer.
Meanwhile, ramping up HBM4 and starting up new fabs will require enormous investment. While pushing ahead with that capital spending, SK hynix also said it aims to return about 100 trillion won to shareholders in fiscal 2026. How far long-term contracts can give the company visibility into future demand, and how it will allocate the profits and cash it generates between capital investment and shareholder returns, are key questions raised by this earnings release.
Revenue Tops 104 Trillion Won, but Net Profit Falls 58% From the Previous Quarter
Third-quarter revenue rose 32% from 79.3187 trillion won in the previous quarter, and operating profit increased 38% from 60.5426 trillion won. Compared with a year earlier, the increases were 328% and 632%, respectively. All figures are provisional consolidated results under Korean International Financial Reporting Standards (K-IFRS) and may change after an audit.
However, it was revenue and operating profit that hit records; net profit declined. Third-quarter net profit was 39.7756 trillion won, down 58% from 93.9226 trillion won in the previous quarter. While operating profit reflects the profitability of the core business, net profit also reflects non-operating gains and losses, taxes, and other items. The announcement did not explain the specific reasons for the drop in net profit, so it cannot be attributed to any particular investment or expense.
SK hynix cited higher prices for key products amid growing demand for AI-related memory, along with increased sales of high-value-added products, as drivers of the revenue and profit growth. In DRAM, sales of server DDR5 rose sharply and HBM4 shipments began in earnest. In NAND, sales of enterprise SSDs (eSSDs) using 321-layer TLC grew, and high-value-added products for data centers performed well. The company did not disclose HBM4's standalone contribution to revenue or profit. It is reasonable to see the strong results as supported by higher sales and prices across multiple products.
Long-Term Supply Contracts Grow From About 10 Customers to 14, and Price Adjustment Mechanism Revealed
SK hynix's long-term supply contracts expanded from about 10 customers to 14 in just three months. The latest announcement also disclosed the share of production capacity covered by the contracts, as well as the mechanism for adjusting sales prices. Comparing the July 29 second-quarter results announcement with the October 28 third-quarter results announcement, the changes in disclosure are as follows.
| Item | July 29 announcement | October 28 announcement |
|---|---|---|
| Number of contracted customers | About 10 | 14 |
| Contracted volume | Not stated in the long-term contract description | Equal to more than 50% of total production capacity for 2027 and 2028 |
| Price adjustment | Not stated in the long-term contract description | Reflects market conditions within an agreed price range |
It is worth noting that the contracted volume shown in October is a share of SK hynix's overall production capacity, not just HBM. Also, the fact that the July announcement did not mention contract volumes or price adjustment details does not necessarily mean such terms did not exist at that point. The latest announcement simply makes the scale of the long-term contracts and the price mechanism easier to grasp.
Securing a certain volume of supply through long-term contracts helps SK hynix gain visibility into medium- to long-term demand. On pricing, however, the contracts reflect market conditions within a range agreed in advance, so they do not uniformly fix future selling prices. The upper and lower bounds of the price range, the products covered for each customer, and the terms if demand declines have not been disclosed. For that reason, the size of the contracted volume alone cannot be taken as a guarantee of future revenue or profit.
According to the company, long-term contracts typically run for five years, and it is also discussing contract extensions with several customers. The customers include major global cloud service providers (CSPs). However, the announcement does not mean buyers have been determined for all production capacity, and supply allocations by product have not been disclosed.
HBM4 to Become the Main Product in Q4 as M15X Capacity Expands
On HBM4, SK hynix gave a specific timeline for its shift to flagship status. In the fourth quarter of 2026, it expects both HBM4 shipments and revenue to exceed those of the previous-generation HBM3E. The outlook is not just for higher shipment volume but for HBM4 to overtake HBM3E in revenue as well, making it the core product of the HBM business.
In its July earnings release, the company said it would begin mass shipments of HBM4 in the second quarter and increase production in the second half. Shipments ramped up in earnest in the third quarter, and the company is now entering a phase of further scaling up production. The M15X fab is at the center of this effort, and the company plans to raise its capacity to the maximum by year-end to increase HBM4 output. However, running M15X at full capacity and HBM4 surpassing HBM3E in both shipments and revenue remain, for now, the company's plans and forecasts.
The production increases are not limited to HBM. In NAND, the company plans to raise the share of its 321-layer process to more than half of its capacity in Korea in the fourth quarter, increasing supply of high-performance, high-capacity enterprise SSDs. Here, "more than half" refers to its domestic capacity, not to a majority of its global capacity. In DRAM, it will accelerate the transition to its sixth-generation 10nm-class (1c) process and expand sales of DDR5 and LPDDR5 SOCAMM for AI servers.
In 2027, the first phase of the Yongin facility is set to begin full-scale operation, and infrastructure investment in Korea and abroad will also expand, so SK hynix expects capital expenditure to rise significantly. The announcement did not give a specific investment figure, however. The company said it will flexibly adjust the scale and timing of investment while taking into account customer demand, product competitiveness, profitability, and investment efficiency.
Behind this aggressive investment is the company's view that the memory supply shortage will persist. Given the technical difficulty of advanced manufacturing processes and the time required to build new fabs, SK hynix forecasts that memory demand will exceed supply through 2030. This is the company's own projection, however, and it is not certain that shortages will continue across the market as a whole. Even with long-term contracts giving visibility into demand, the challenge remains of bringing facilities online as planned and supplying the required products.
About 100 Trillion Won in Shareholder Returns, Excluding Share Buybacks for Employee Compensation
To consider how SK hynix can balance heavy capital spending with shareholder returns, it is necessary to distinguish between profit booked in a quarter and cash actually on hand. At the end of the third quarter, cash and cash equivalents stood at 114.5 trillion won and borrowings at 15.8 trillion won, leaving net cash of 98.7 trillion won. At the end of the previous quarter, cash and cash equivalents were 88 trillion won, borrowings were 18.6 trillion won, and net cash was 69.4 trillion won. These figures represent balances of cash and debt at the end of each quarter, not operating or net profit booked over a period.
For fiscal 2026 shareholder returns, SK hynix said it aims for a total of about 100 trillion won, built around cash dividends of at least 30 trillion won and share repurchases and cancellations of at least 60 trillion won. The repurchases and cancellations include the 40 trillion won announced in August. The company is working toward its policy of returning at least 50% of free cash flow (FCF) to shareholders. However, about 100 trillion won is a full-year target, not a sum that has already been paid out. The 40 trillion won buyback announced in August was completed last week, but cancellation of the acquired shares is scheduled for a later date.
At its October 28 board meeting, the company approved a third-quarter cash dividend of 14 trillion won and a new 20 trillion won share buyback. However, the entire 20 trillion won buyback does not count as shareholder return. Of that amount, 10 trillion won will be acquired for shareholder returns and fully cancelled, while the remaining roughly 10 trillion won will be used for employee compensation and excluded from total shareholder returns. It would therefore be inaccurate to simply add the 14 trillion won dividend and the 20 trillion won buyback and treat 34 trillion won as the shareholder return amount this time.
The third-quarter dividend per share is 20,000 won, consisting of a regular dividend of 10,000 won reflecting improved earnings and a special dividend of 10,000 won distributing gains from its investment in Kioxia to shareholders. SK hynix's basic policy is to return most of the profits from strategic investments to shareholders. Following that policy, it intends to return most of this year's realized Kioxia investment gains through the special dividend.
On the other hand, the company changed its earlier plan for funding capital expenditure. CEO Kwak Noh-jung explained that the original purpose of Solidigm's initial public offering (IPO) was to raise the outside funds needed to expand production facilities, easing SK hynix's own investment burden and widening its capacity for shareholder returns. After reviewing financial market conditions and other factors, however, the company concluded it could not fully achieve the effect it had originally hoped for and halted its consideration of the IPO. This is a change in funding policy and does not mean that capital investment in Solidigm is no longer needed.
In addition to record revenue and operating profit, these results revealed an expansion of long-term supply contracts and concrete shareholder return measures. The focus now is whether HBM4 actually surpasses HBM3E in both shipments and revenue, and whether the M15X and Yongin production facilities translate into expanded supply capacity as planned. Once the scale and execution of investment from 2027 onward become clear, it will be easier to see how far the demand secured through long-term contracts turns into actual revenue and profit, and how SK hynix balances capital spending with large-scale shareholder returns.