To The Moon

$SK hynix(SKHY)$ — Is This the Ultimate Answer to the AI Bubble Debate?

Record earnings are strengthening the case for AI infrastructure spending. But for investors, the bigger question is whether today's extraordinary profitability can last.

SK hynix has delivered a quarterly earnings report that is difficult to ignore.

On October 28, the South Korean memory chipmaker reported third-quarter 2026 revenue of ₩104.62 trillion, up 328% year over year, while operating profit surged 632% to ₩83.33 trillion. Both figures set new quarterly records, and revenue crossed the ₩100 trillion threshold for the first time in the company's history.

Behind these numbers lies a fundamental shift in the semiconductor industry: memory is becoming an increasingly critical component of AI infrastructure, rather than simply a cyclical commodity used in conventional computing.

But record earnings do not automatically mean a stock is undervalued.

For investors, three questions matter most: Is AI demand translating into sustainable profits? Can SK hynix maintain its competitive advantage as the HBM market evolves? And how much of this growth is already reflected in its valuation?

1. Record Results: The Numbers Behind the AI Boom

SK hynix's Q3 results demonstrate the extraordinary operating leverage that can emerge when strong demand meets constrained supply.

Q3 2026 Financial Highlights

Source: SK hynix's Q3 2026 earnings release. Figures are based on consolidated K-IFRS financial statements.

The operating margin is particularly striking. SK hynix generated nearly ₩80 in operating profit for every ₩100 of revenue.

This reflects more than an increase in shipment volumes. The company is benefiting from higher memory prices, a richer product mix and strong demand for high-value-added products, including high-bandwidth memory (HBM) and server-grade DDR5.

Revenue increased 32% quarter over quarter, while operating profit rose 38%, indicating that profitability continued to improve alongside sales.

The results also exceeded market expectations. Reported operating profit came in above the approximately ₩76.96 trillion consensus cited by Korean financial media.

The key takeaway: SK hynix is capturing substantial economic value from the AI infrastructure buildout. The next challenge is maintaining that profitability as capacity expands and competitors introduce new products.

2. Why HBM Is Changing the Economics of the Memory Industry

The most important driver behind SK hynix's performance is the growing importance of memory in AI computing.

Training and running advanced AI models requires more than powerful processors. It also requires rapid access to enormous amounts of data.

High-bandwidth memory addresses this challenge by providing very high data-transfer bandwidth in a compact package designed to work closely with advanced computing processors.

As AI accelerators become more powerful, memory performance can become a critical factor in overall system performance.

That creates a valuable opportunity for suppliers capable of delivering advanced HBM products at scale.

Three Drivers Behind SK hynix's Growth

1. High-bandwidth memory

HBM has become a strategic component of advanced AI accelerators. SK hynix has established a strong position in this market, benefiting from its technical capabilities, customer relationships and experience scaling production.

2. Server DRAM

AI infrastructure requires substantial memory capacity across servers and computing systems. Demand for high-performance DDR5 and other server-oriented memory products is providing an additional growth engine beyond HBM.

3. Enterprise SSDs

AI data centers need efficient storage systems to manage expanding datasets, model checkpoints and inference workloads. SK hynix reported growth in sales of high-value-added enterprise SSDs, including products based on 321-layer NAND technology.

These businesses reinforce one another.

As AI infrastructure expands, demand can increase across accelerators, system memory and enterprise storage. SK hynix is therefore positioned to benefit from multiple layers of the AI hardware ecosystem.

However, investors should distinguish between structural growth in AI computing and the profitability of any individual supplier. A growing market does not guarantee that today's margins will remain unchanged.

3. HBM4: The Next Test of SK hynix's Competitive Advantage

The transition from HBM3E to HBM4 is one of the most important developments to watch.

SK hynix expects HBM4 to become its mainstream HBM product in Q4 2026, surpassing HBM3E in both shipment volume and revenue. The company is also accelerating production expansion at its M15X facility to support growing demand.

This transition matters for three reasons.

First, product leadership must translate into commercial shipments.

Developing advanced memory is only part of the challenge. Manufacturers must also achieve reliable production yields, meet customer qualification requirements and deliver sufficient volumes on schedule.

Second, customers need dependable supply.

AI accelerator manufacturers and cloud service providers plan infrastructure deployment well in advance. Suppliers that can deliver qualified products consistently may benefit from stronger customer relationships and greater revenue visibility.

Third, the competitive landscape is evolving.

Samsung Electronics and Micron are also competing for opportunities in advanced HBM. SK hynix's existing position is valuable, but it cannot be treated as permanent.

The investment question is not simply whether HBM4 demand will grow.

It is whether SK hynix can preserve its technological lead, production efficiency and pricing power as competitors improve their offerings.

If it can, the HBM4 transition could support another stage of earnings growth. If competition intensifies faster than expected, rising industry volumes may not translate into equivalent profit growth.

4. The Strongest Evidence Against the AI Bubble Argument: Forward Visibility

One of the most significant details in SK hynix's announcement is its progress on long-term customer agreements.

The company said it had signed long-term agreements with 14 major customers, including leading global cloud service providers. These agreements typically run for five years, and the company is discussing possible extensions with several customers.

More importantly, the contracted volume covers more than 50% of total production capacity for 2027 and 2028.

This is meaningful because it gives SK hynix greater visibility into future demand than a business relying entirely on short-term orders.

The agreements also incorporate pricing mechanisms that reflect market conditions within ranges agreed upon with customers. According to the company, this structure is intended to improve demand visibility while supporting stable profitability.

For investors evaluating the AI bubble debate, this is an important distinction.

A speculative narrative is one thing. Customers committing to substantial future supply is another.

Long-term agreements suggest that major customers are planning for continued memory demand rather than treating AI infrastructure as a short-lived experiment.

However, there is an important qualification: contracted production capacity is not the same as guaranteed future profit.

The final outcome will still depend on actual deliveries, customer demand, production costs, agreed pricing and the financial health of the wider AI ecosystem.

Long-term agreements reduce uncertainty; they do not eliminate it.

5. Is the AI Bubble Narrative Dead?

Not yet.

SK hynix's results provide strong evidence that AI-related investment is generating real revenue and substantial profits for at least some infrastructure suppliers.

But the AI bubble debate involves two separate questions that investors should not confuse.

Question A: Is AI Demand Real?

The evidence is increasingly tangible.

SK hynix's record revenue, rising operating profit and expanding sales of HBM and server memory demonstrate that AI-related demand is translating into commercial activity.

The company's long-term customer agreements provide further evidence of forward planning among major buyers.

This weakens the argument that the entire AI infrastructure boom is based on hype without meaningful underlying demand.

Question B: Are AI-Related Stocks Correctly Valued?

That is a different question.

Even a company with outstanding earnings can become an unattractive investment if its share price already assumes years of uninterrupted growth.

Investors must consider the relationship between future earnings, valuation multiples, capital expenditure and the durability of current margins.

For SK hynix, the critical issue is whether today's extraordinary profitability represents a sustainable earnings base or a peak in a particularly favorable memory cycle.

Real demand does not automatically justify every valuation. Real profits do not eliminate cyclical risk.

The strongest investment case combines genuine structural demand with a reasonable entry price and evidence that the company can defend its competitive position.

6. The Risk Investors Should Not Ignore: An 80% Operating Margin Is an Exceptionally High Bar

SK hynix's nearly 80% operating margin is impressive, but it also creates a difficult comparison for future quarters.

When profitability reaches exceptional levels, even a modest deterioration in pricing or product mix can have a meaningful effect on earnings.

Several risks deserve attention.

Memory pricing could normalize.

If supply expands faster than demand, pricing power could weaken. Because memory is a cyclical industry, periods of tight supply can eventually give way to more balanced market conditions.

Competition could intensify.

Rivals are investing in advanced HBM and other high-performance memory products. Greater competition could affect market share, pricing and customer relationships.

AI infrastructure spending could slow.

Cloud providers and other large customers must eventually generate sufficient economic returns from their AI investments. If monetization disappoints or infrastructure budgets are reassessed, orders could slow.

Capital expenditure could rise.

SK hynix plans to expand production capacity to capture future demand. That investment may support long-term growth, but it also increases the importance of capacity utilization, execution and returns on invested capital.

Valuation could remain vulnerable.

Markets often price semiconductor stocks based on expectations several quarters or years into the future. A company can report record results and still see its share price fall if investors expected even stronger performance.

These risks do not negate the bullish earnings story. They explain why investors should evaluate the durability of the earnings rather than simply extrapolate the latest growth rate.

7. What Investors Should Watch in Q4 2026 and 2027

The next stage of the story will be determined by execution.

The most useful signal will be the relationship between revenue growth and operating profitability.

If sales continue to rise while margins remain strong, the market may gain confidence that the company is benefiting from structural changes in AI computing rather than only a temporary pricing upswing.

If revenue grows but margins contract sharply, investors will need to reassess how much of the current earnings surge can be sustained.

8. Final Verdict: A Strong AI Infrastructure Business, but Not a Risk-Free Investment

SK hynix's Q3 2026 results make a compelling case that AI infrastructure demand is creating real economic value.

Quarterly revenue exceeded ₩100 trillion for the first time. Operating profit reached ₩83.33 trillion. HBM4 is moving toward becoming the company's mainstream HBM product, while long-term agreements cover more than half of its production capacity for 2027 and 2028.

Together, these developments support the view that AI memory demand has a substantial commercial foundation.

But the investment conclusion should remain balanced.

The bullish case: SK hynix has a strong position in a strategically important segment of AI infrastructure, rising demand across multiple memory categories and improved visibility into future production.

The cautious case: Current margins are exceptionally high, competition is intensifying, memory remains cyclical, and the stock's attractiveness depends on valuation as well as business performance.

The next phase is not simply about whether AI demand continues to grow. It is about whether SK hynix can convert that growth into sustainable profits and attractive returns on the capital it invests.

Bottom line: SK hynix has delivered powerful evidence that AI memory demand is real. Whether its shares offer an attractive investment opportunity now depends on the price investors pay, the durability of its competitive advantage and the company's ability to defend profitability as the industry expands.

The AI bubble debate is far from settled. But for SK hynix, the discussion is increasingly grounded in real revenue, real customer commitments and real earnings.

Join the Discussion

Which factor will matter most for SK hynix over the next 12 months?

• A. HBM4 adoption and market-share leadership

• B. Memory pricing and the sustainability of current margins

• C. AI infrastructure spending by major cloud providers

• D. Valuation and the risk of a semiconductor cycle reversal

Share your view in the comments. The most important question may not be whether AI demand is real, but how much of that future growth is already priced into the stock.

Disclaimer: This article is for informational and educational purposes only and does not constitute investment advice or a recommendation to buy or sell any security. Financial results are based on SK hynix's Q3 2026 announcement dated October 28, 2026; figures and forward-looking expectations may change.