Samsung Electronics just posted its best quarterly profit in more than two years. The reverberations are being felt thousands of miles away in Boise, Idaho.
Micron Technology shares surged more than 4% in premarket trading on Thursday after Samsung reported operating profit of 6.61 trillion Korean won ($4.63 billion) for the second quarter — a figure that obliterated analyst expectations and marked a staggering turnaround from the 4.36 trillion won loss the Korean giant posted in the same period a year ago. The signal to memory chip investors was unmistakable: the cyclical recovery in DRAM and NAND flash memory isn’t just holding. It’s accelerating.
As Barron’s reported, Samsung’s preliminary earnings release — the company provides guidance ahead of its full earnings report — showed revenue of approximately 74 trillion won, up roughly 23% from the prior quarter. That sequential jump is enormous by any standard, and it reflects surging demand for high-bandwidth memory chips used in artificial intelligence servers, alongside a broader recovery in conventional memory pricing that had been in a brutal downturn for much of 2023.
The read-through to Micron is direct and powerful. Micron, the last major U.S.-headquartered memory chipmaker, derives the vast majority of its revenue from DRAM and NAND products that compete head-to-head with Samsung and SK Hynix. When Samsung reports blowout numbers driven by memory pricing strength and AI-related demand, Micron’s own trajectory tends to follow. Wall Street knows this. Hence the stock reaction.
But there’s a deeper story here than a simple sector rotation trade.
The memory chip industry has historically been one of the most brutally cyclical corners of the semiconductor world. Boom follows bust with a regularity that punishes latecomers and rewards those who understand the cadence. The 2022-2023 downturn was severe — Samsung posted consecutive quarterly losses in its semiconductor division, Micron slashed capital expenditures, and SK Hynix cut production. Inventory gluts crushed pricing. The AI boom, catalyzed by the explosive adoption of large language models and generative AI applications, changed the equation almost overnight.
What’s different this time is the composition of demand. Traditional memory cycles were driven by PCs and smartphones. Those markets still matter, but the incremental demand engine is now data center infrastructure — specifically, the high-bandwidth memory (HBM) chips required to feed Nvidia’s GPU clusters. HBM is a premium product with significantly higher margins than commodity DRAM. SK Hynix has led the HBM market, but Samsung has been investing heavily to close the gap, and Micron has been shipping its own HBM3E product to major customers including Nvidia.
Samsung’s results suggest the HBM ramp is proceeding faster than many analysts expected. According to Barron’s, the Korean company’s semiconductor division likely accounted for the bulk of the profit improvement, with memory chips doing most of the heavy lifting. Samsung doesn’t break out HBM revenue specifically in its preliminary release, but industry analysts have estimated that HBM could represent a mid-teens percentage of total DRAM industry revenue by the end of 2024, up from low single digits just two years ago.
For Micron, the implications are layered. The company reported its own strong fiscal third-quarter results in late June, with revenue of $6.81 billion — up 82% year over year — and guidance that exceeded expectations. CEO Sanjay Mehrotra told analysts that Micron’s HBM production was sold out for 2024 and that the “overwhelming majority” of 2025 supply was already allocated. The stock rallied on those results but has since pulled back amid a broader rotation out of AI-related names.
Samsung’s numbers could reignite that trade.
There’s a valuation argument to consider, too. Micron shares trade at roughly 11 times forward earnings, a discount to the broader semiconductor index and a fraction of the multiples commanded by AI darlings like Nvidia and Broadcom. Memory chip stocks have always traded at lower multiples because of their cyclicality, but some analysts argue the AI-driven structural demand shift warrants a rerating. If HBM and data center DRAM represent a more durable, less cyclical revenue stream than the PC and smartphone markets that dominated previous cycles, perhaps Micron deserves a higher multiple than history suggests.
Not everyone agrees. Bears point out that memory chip companies have a long track record of overinvesting at the top of the cycle, leading to supply gluts that destroy pricing power. Samsung, SK Hynix, and Micron are all ramping capital expenditures aggressively right now. Samsung alone has committed to building new advanced memory fabrication capacity in South Korea and is expanding its facility in Taylor, Texas. If AI spending plateaus or the macro environment deteriorates, the industry could find itself with too much supply chasing too little demand. Again.
The geopolitical dimension adds another variable. China’s memory chip industry, led by CJMT (ChangXin Memory Technologies) and YMTC (Yangtze Memory Technologies), has been constrained by U.S. export controls on advanced chipmaking equipment. Those restrictions have effectively frozen Chinese competitors out of the most advanced DRAM and NAND nodes, giving Samsung, SK Hynix, and Micron breathing room at the high end of the market. But the restrictions are a policy choice, not a law of physics. Any relaxation — or successful Chinese workaround — could alter the competitive dynamics significantly.
For now, the momentum is clearly with the incumbents. And Samsung’s second-quarter results are the latest data point confirming that the memory upcycle has legs.
SK Hynix, which reports its own preliminary results soon, is expected to show a similarly dramatic profit recovery. The company has been the primary beneficiary of the HBM boom, having secured early design wins with Nvidia for its HBM3 and HBM3E products. If SK Hynix’s numbers come in as strong as Samsung’s, the bullish case for the entire memory sector — Micron included — gets even harder to dismiss.
There’s also the question of what Samsung’s results mean for the broader AI supply chain narrative. Nvidia’s next-generation Blackwell GPU platform, expected to ramp in the second half of 2024 and into 2025, will require even more HBM per chip than the current Hopper architecture. That means memory content per server is going up, not down. Every major cloud hyperscaler — Microsoft, Google, Amazon, Meta — is building out AI infrastructure at a pace that shows no sign of slowing. The capital expenditure commitments from these companies for 2024 and 2025 are staggering, collectively running into hundreds of billions of dollars. A meaningful portion of that spend flows directly to memory chip suppliers.
Micron is positioning itself to capture a growing share of that spend. The company recently broke ground on a new fabrication facility in Clay, New York, supported by up to $6.1 billion in CHIPS Act grants from the U.S. government. That facility, along with expansions in Boise and Manassas, Virginia, will focus on advanced DRAM production, including HBM. It’s a multi-year buildout, but it signals Micron’s intent to compete aggressively at the leading edge of memory technology.
The stock market’s reaction to Samsung’s earnings tells you something about where sentiment stands. Investors are hungry for confirmation that the AI spending cycle is translating into real, measurable profits for semiconductor companies beyond Nvidia. Samsung just provided that confirmation. Micron, as the most direct U.S.-listed proxy for the memory chip trade, benefits immediately.
Whether that benefit is sustainable depends on execution, pricing discipline, and the durability of AI-related demand. Those are real uncertainties. But the direction of the cycle, right now, is unambiguous.
Up.
WebProNews is an iEntry Publication