Samsung's Pricing Power

The "Boring" Memory Chip

Just a few months ago, the story of the AI memory market was simple: SK Hynix was the new king, and Samsung, the long-reigning monarch, had been humbled. Having masterfully aligned itself with Nvidia, SK Hynix became the dominant supplier of High Bandwidth Memory (HBM). Samsung, meanwhile, stumbled, struggled to pass Nvidia's quality tests, and saw its market share plummet. The crown had been passed.

Or so it seemed.

Last week, in a confident move, Samsung hiked the price of its memory chips by as much as 60%. Not a gentle nudge, but a massive, take-it-or-leave-it price jack that, according to Reuters, has customers "panic buying" and is sending shockwaves through the entire tech supply chain:

The chip crunch has been so severe that it has spurred panic buying by some customers, according to industry executives and analysts.

China's top contract chipmaker SMIC said on Friday that the memory chip shortage has meant that customers are holding back orders for other types of chips that are also used in their products.

Xiaomi, a Chinese smartphone, electronics and auto manufacturer, also warned last month that the surging prices have raised the cost of making phones.

This is the move of a dominant player reasserting his power. So what changed? How did the company that was on the ropes just a few months ago suddenly find the leverage to dictate terms to the entire industry?

The answer is a two-part strategy of brutal opportunism and a hard-fought comeback.

The first part is about the boring chips. While the entire world was obsessed with the high-stakes race for HBM, the massive AI build-out was creating a crippling shortage of the unglamorous memory: standard DDR5 that goes into every single server. Because SK Hynix was dedicating so much of its production to HBM, it created a supply vacuum. Samsung, by virtue of being slower to pivot its factories, found itself with a near-monopoly on a suddenly scarce and incredibly valuable resource. As a recent report noted, the market for general-purpose DRAM has become so tight that analysts expect operating margins for key suppliers to top 70%, a level not seen since the PC boom of 1995.

But this isn't just a story of an accidental windfall. While exploiting its dominance in the commodity market, Samsung was also staging a furious comeback in the high-end HBM race. The proof of this turnaround arrived last month in its third-quarter earnings report. It was a blockbuster, with the company posting its highest-ever quarterly revenue and operating profit surging back into the double-digit trillions of won for the first time in three years.

The driver, according to the company, was a perfect storm of success. Its results were powered by brisk sales of HBM chips—signaling its comeback at the high end—combined with strong demand for the very DDR5 chips that are now in short supply.

This financial turnaround was powered by real progress: the company reportedly passed Nvidia's long-awaited quality tests, began mass-producing its advanced HBM3E chips, and, according to a recent earnings call, has already sold out its entire 2026 HBM supply.

This is the context for last week's price hike. Samsung now finds itself in an almost unbelievably dominant position: it has a chokehold on the supply of the "boring" commodity chips everyone needs, and it is closing the gap on the high-end strategic chips everyone wants.

For months, the market wondered if Samsung could still compete. Now, it's not just competing; it's dictating prices.

More on AI Memory:

• Can Google and Meta Execs Fix AI Data Centre Memory Limits? (Data Centre Magazine)
• Is AI About to Hit a 'Memory Wall'? (ARPU)

On Our Radar

Our Intelligence Desk connects the dots across functions—from GTM to Operations—and delivers intelligence tailored for specific roles. Learn more about our bespoke streams.

Oracle's Hundred-Billion-Dollar Gamble

• The Headline: Oracle Bonds Sell Off Amid Investor Concern Over its Debt-Fueled AI Infrastructure Strategy (Reuters)
• ARPU's Take: This is the market showing its nervousness about Oracle's attempt to buy its way into the top tier of the AI cloud race. With over $100 billion in existing debt and reports of another $38 billion planned, the market is questioning the financial sustainability behind Oracle's capital expenditure strategy.
• The Operations Implication: This bond sell-off highlights the fundamental operational difference between Oracle and the top-tier hyperscalers. While rivals like Microsoft and Google can fund their massive AI capex primarily through free cash flow from their core businesses, Oracle has already flipped it free cash flow negative (-$6 billion TTM) and is forced to take on substantial debt. A $100 billion-scale, multi-year infrastructure commitment would require Oracle to sustain capex at levels at least 5 times its pre-AI norms for years, almost entirely debt-financed at a time when its total debt already exceeds $100B. This makes its AI strategy inherently riskier and more vulnerable to credit market sentiment.

The AI Hype Check

• The Headline: RBC Analysts Flag Enterprise AI Slowdown, Citing a 'Productivity Paradox' and 'Pilot Fatigue' (Business Insider)
• ARPU's Take: The initial sugar rush of enterprise AI adoption is over. Now comes the hard part: RBC analysts are flagging what they call a "productivity paradox"—the promised efficiency gains aren't showing up, leading to "pilot fatigue" as budget holders start asking for the receipts.
• The Go-to-Market Implication: This signals the AI market is entering a more challenging 'prove it' phase of the adoption cycle. The initial wave of enterprise adoption, driven by experimentation and FOMO, is giving way to a more pragmatic phase where buyers demand clear, quantifiable ROI. For AI service providers, this means the go-to-market strategy must shift from selling generic platform capabilities to selling pre-packaged, vertical-specific solutions that solve a tangible business problem with a clear financial return. This puts pressure on vendors like Microsoft, Google, and OpenAI to move beyond API access and demonstrate concrete, out-of-the-box value.

P.S. Tracking these kinds of complex, cross-functional signals is what we do. If you have a specific intelligence challenge that goes beyond the headlines, get in touch to design your custom intelligence.

You received this message because you are subscribed to ARPU newsletter. If a friend forwarded you this message, sign up here to get it in your inbox.