SK hynix could surpass Microsoft and Google in profitability amid AI memory boom
SK hynix could become one of the main financial beneficiaries of the AI boom. Analysts expect operating profit growth of 432% in 2026 — reaching $167…
AI-processed from 3DNews AI; edited by Hamidun News
SK hynix may become one of the most notable financial winners of the artificial intelligence boom in 2026. Analysts expect the company to sharply increase operating profit thanks to demand for AI-server memory and maintain high earnings even as Samsung advances in HBM4.
2026 Forecast
According to KB Securities estimates, SK hynix's operating profit in 2026 could grow by 432% and reach $167 billion. If this scenario materializes, the South Korean memory manufacturer will rise to the ranks of the world's most profitable companies and surpass Microsoft and Google individually by this metric.
For the chip market, this is an unusual moment: until recently, memory manufacturers were considered a cyclical business with sharp downturns, but now they are becoming among the main beneficiaries of the AI race. The reason is that AI infrastructure requires enormous volumes of expensive and technologically complex memory. This is not only about HBM, which is used alongside accelerators, but also about classical DRAM and NAND, whose demand is growing alongside data center expansion.
Analysts expect SK hynix's profit to increase throughout 2026, rather than as a one-time spike. This makes the forecast particularly important: the market is betting not on short-term hype, but on a sustainable cycle of demand.
Growth Drivers
The main engine is the price increase across almost the entire memory lineup against the backdrop of overheated demand from AI and cloud infrastructure providers. The stronger prices rise on chips as production costs grow moderately, the higher the operating margin. For SK hynix, this is especially advantageous because the company has already established itself as a key supplier of HBM for accelerators, meaning it can sell not just more memory, but more expensive and profitable products.
In the 2026 forecast, this is expressed in several illustrative figures:
- operating profit: +432% year-over-year
- profit forecast: approximately $167 billion
- expected DRAM price increase: +170%
- expected NAND price increase: +190%
- 2027 forecast: up to $241 billion
The logic itself is simple: operating profit is the difference between revenue and expenses. If memory manufacturers sell products at higher prices than their costs grow, the financial leverage works very quickly. This is why the forecasts feature such large values, which until recently seemed impossible for a semiconductor company specializing primarily in memory.
For SK hynix, the product mix is also important: the higher the share of HBM and other scarce solutions, the stronger the overall profitability grows.
Samsung Factor
A separate question is whether Samsung Electronics can weaken its competitor's position if it accelerates HBM4 adoption. So far, analysts do not consider this a decisive risk for SK hynix. The logic is that the AI memory market is growing so rapidly that Samsung's progress alone should not automatically deprive SK hynix of high profitability. Moreover, the very fact of increased competition does not eliminate the shortage of advanced components for AI systems, meaning it preserves space for high prices and large contracts.
However, in a longer-term horizon, competition will only intensify. According to forecasts, already in 2027, SK hynix's operating profit could grow to $241 billion, which would place the company in third place globally by this metric. Samsung Electronics is expected to be in first place, which also earns from memory and other areas of electronics.
For investors and the entire market, this is a signal that the center of power in the AI chain is shifting from model developers alone to suppliers of key hardware components.
What This Means
The AI boom brings excess profits not only to model developers and cloud platforms, but also to companies that supply the basic infrastructure for their operation. The story of SK hynix shows that memory has become one of the most expensive and scarce resources in the new AI economy, which means that competition for HBM, DRAM and NAND will directly affect both the chip market and the balance of power among the largest technology companies.
Want to stop reading about AI and start using it?
AI News is a curated feed of AI/tech news. Hamidun Academy teaches you to use AI systematically in your work.