#股票财报季 SK Hynix’s Q3 earnings are imminent: profits may break records again, but HBM4 has become a “semi-finished” engine
On October 28, SK Hynix will release its third-quarter 2026 results. Market consensus expects revenue of about 98 trillion won and operating profit of about 77 trillion won, with an operating margin approaching 78%—if realized, it would mark the sixth consecutive quarter of record highs.
On the surface, the AI memory cycle remains疯狂, with price increases in general-purpose DRAM and existing HBM demand pushing profits to historic highs. But the real focus is HBM4: the company began mass-production shipments in Q2, and shipments should have accelerated in the second half of the year, yet Korean media and industry sources indicate that Q3 shipments fell short of expectations for the second consecutive quarter. The delay in Nvidia’s Rubin platform + the time required for production tuning have caused part of the capacity to shift to traditional DRAM.
More importantly, the competitive landscape is changing. Bernstein and other institutions estimate that Samsung’s Q3 HBM revenue may have already surpassed SK’s, with the share gap narrowing rapidly from more than 30 percentage points. SK’s first-mover advantage still remains, but the situation of “one company dominating alone” has begun to loosen.
So this quarter’s record performance is being supported more by older momentum, while the real boost from HBM4 depends on the specific numbers disclosed in the conference call: shipment pace, progress on long-term customer contracts, and 2027 capacity plans. High profits are a fact, but the market wants to confirm more—whether the next growth engine is truly stable.
On October 28, SK Hynix will release its third-quarter 2026 results. Market consensus expects revenue of about 98 trillion won and operating profit of about 77 trillion won, with an operating margin approaching 78%—if realized, it would mark the sixth consecutive quarter of record highs.
On the surface, the AI memory cycle remains疯狂, with price increases in general-purpose DRAM and existing HBM demand pushing profits to historic highs. But the real focus is HBM4: the company began mass-production shipments in Q2, and shipments should have accelerated in the second half of the year, yet Korean media and industry sources indicate that Q3 shipments fell short of expectations for the second consecutive quarter. The delay in Nvidia’s Rubin platform + the time required for production tuning have caused part of the capacity to shift to traditional DRAM.
More importantly, the competitive landscape is changing. Bernstein and other institutions estimate that Samsung’s Q3 HBM revenue may have already surpassed SK’s, with the share gap narrowing rapidly from more than 30 percentage points. SK’s first-mover advantage still remains, but the situation of “one company dominating alone” has begun to loosen.
So this quarter’s record performance is being supported more by older momentum, while the real boost from HBM4 depends on the specific numbers disclosed in the conference call: shipment pace, progress on long-term customer contracts, and 2027 capacity plans. High profits are a fact, but the market wants to confirm more—whether the next growth engine is truly stable.