SK Hynix — one of the world's largest memory-chip makers — has reported record third-quarter 2026 results, with revenue of about 17.6 trillion won, up as much as 94% from a year earlier. The main driver is demand for memory chips used in artificial intelligence, especially from data centers. This isn't just one company's news: it partly explains why RAM and memory prices are climbing, which feeds into the cost of many tech devices.

Why does AI need so much memory?

Modern AI models need enormous amounts of memory to train and run. In particular, high-bandwidth memory (HBM) — stacked in many layers to move data extremely fast — is a key component in the machines that run AI in data centers. As tech companies race to build AI infrastructure, they buy HBM and memory chips in huge volumes, pushing demand very high. SK Hynix and Samsung are top suppliers, so they benefit directly.

  • SK Hynix: record Q3 revenue of ~17.6 trillion won, up ~94% year over year.
  • Driver: AI memory demand, especially high-bandwidth HBM for data centers.
  • Samsung also benefited strongly from memory chips in the same period.

How does AI demand push up RAM prices?

When much of memory production and supply is channeled to HBM and AI customers, the ordinary memory (DRAM) available for computers, phones and consumer devices gets tighter, pushing prices up. Per market reports, server DRAM prices were expected to rise by double-digit percentages in this period. Higher component prices tend to seep into final product prices: computers, phones, and RAM/storage upgrades. The effect isn't instant — device makers often buy memory in advance and absorb some cost — but over months it shows up as pricier configurations, smaller discounts, or base models that ship with less storage than you'd like. It's one of the less visible ways the AI boom reaches ordinary shoppers, long before most people ever use an AI data center directly.

What does it mean for the device you want to buy?

This is the thread explaining why some recent high-end tech products cost more. As companies ship new machines with more RAM and storage, high component costs make low prices hard to offer. In other words, the AI chip boom at the data-center level is echoing all the way down to the wallets of people buying computers and phones. One practical takeaway: if you plan to upgrade RAM or storage, rising prices may be a reason to think it over more carefully.

The HBM race: who's leading?

The market for high-bandwidth memory (HBM) — the priciest and most important component for AI — is a fierce race. SK Hynix has led for years with a large share, while Samsung has accelerated and is reported to have gained significant share after launching new HBM generations. America's Micron is also in the mix. Because HBM brings high margins and demand that supply can barely meet, makers are pouring production capacity into it — which is exactly why ordinary consumer memory is becoming scarcer and more expensive.

What should buyers do?

For people buying a computer or phone, the practical advice is to pick a configuration that's enough for your real needs, rather than chasing the highest RAM/storage tier while prices are high. If you plan to upgrade RAM or an SSD yourself, you can watch prices for a while and buy on a deal, since component prices move in cycles. Conversely, if you genuinely need lots of memory for work (video editing, virtual machines, large datasets), buying sooner can sometimes beat waiting — because the broad trend in this period is prices edging up, not down.

Caveats on the numbers

  • Revenue/profit figures can differ across sources — treat the company's official reports as the standard.
  • Rising RAM prices are a broad trend driven by AI demand, but the exact increase varies by chip type and timing.
  • The memory-chip cycle rises and falls; prices may cool when supply catches up, though it's unclear when.

A boom with two sides

For memory makers, the AI boom is a windfall after years of ups and downs in a notoriously cyclical business. Record revenue and strong demand let them invest in new factories and advanced chips. But the same boom has a cost for ordinary buyers: when the most profitable chips go to AI customers, consumer memory gets tighter and pricier, nudging up the cost of everyday computers and phones. It's a reminder that the AI story isn't only about chatbots and data centers — it quietly reaches the price tag on the device in your hand.

How long this lasts is the open question. Memory has always been cyclical: when prices are high, makers expand capacity, supply eventually catches up, and prices fall again. The twist this time is how large and sustained AI demand looks, which could keep the cycle elevated longer than usual. For now, the sensible reading is that memory-related prices are more likely to stay firm than to drop sharply in the near term — useful context whether you're buying a laptop, a phone, or just a RAM upgrade.

FAQ

  • What did SK Hynix report? — Record Q3 2026 revenue of ~17.6 trillion won, up ~94% year over year.
  • Why the big result? — A boom in AI memory demand, especially HBM for data centers.
  • What is HBM? — High-bandwidth memory, stacked in layers, used in AI machines.
  • Why are PC/phone RAM prices rising? — Supply is channeled to AI, leaving consumer DRAM tighter and pricier.
  • Will prices come down? — The memory cycle rises and falls; prices may ease when supply catches up, timing unclear.

This article is part of DETOON TODAY, compiled from publicly reported details. Sources: Korea JoongAng Daily, Sammy Fans (2026).