To The Moon

After the U.S. market closed, $希捷科技(STX)$ delivered a major earnings beat that showed AI storage demand remains strong. Just a few hours later, $SK海力士(SKHY)$reported record results that still fell short of expectations, sending a very different message: strong demand alone is no longer enough.

Here is the main takeaway:

The fundamentals of the storage industry have not suddenly reversed, but the way the market values the sector has changed. Previously, rising prices and growing orders were enough to lift the entire group. Now companies must beat expectations by a wide margin while also convincing investors that future capacity expansion will not create another supply glut.

Seagate clearly passed that test.

The company reported quarterly revenue of $3.63 billion, up 48.5% year over year. Adjusted EPS reached $5.71, non-GAAP gross margin rose to 52.7%, and free cash flow came in at $1.12 billion.

Its guidance was even stronger.

Seagate expects next-quarter revenue of approximately $4.1 billion, well above the market’s previous estimate of around $3.75 billion. Adjusted EPS is expected to reach $7.30, compared with consensus expectations of roughly $5.80.

This was not an ordinary earnings beat. Revenue, margins, and earnings are all expected to accelerate again. STX jumped more than 8% in after-hours trading, recovering most of its losses from the regular session.

Seagate proved one important point: AI-generated data continues to grow rapidly, and hyperscalers have not stopped buying high-capacity storage.

SK Hynix, however, told a different story.

The company reported second-quarter revenue of KRW 79.32 trillion, up 51% sequentially, while operating profit rose 61% to KRW 60.54 trillion. Its operating margin reached a record 76%.

The problem was that the market had expected approximately KRW 84.06 trillion in revenue and KRW 64.09 trillion in operating profit. Both figures missed consensus estimates by roughly 5.5%.

Under normal conditions, this would still be an extremely strong earnings report. But SK Hynix had already fallen 14.7% the previous day, and investors were counting on a huge beat to rescue the entire storage sector.

Instead, they got record results that were still not good enough.

What collapsed was not storage demand, but the expectation of an immediate V-shaped recovery driven by SK Hynix’s earnings.

The company’s operating data did not show any real demand breakdown.

DRAM shipments increased by a high-single-digit percentage sequentially, while average selling prices rose by roughly 30%. NAND shipments grew by a mid-teens percentage, with average selling prices rising by approximately 50%.

For the third quarter, SK Hynix expects DRAM shipments to grow by another 10%, while NAND shipments are expected to increase by a low-single-digit percentage.

HBM4 shipments have already begun and will ramp further during the second half of the year. HBM4E samples have been delivered to customers, while SK Hynix has completed long-term supply agreements with around ten clients, with some contracts including prepayment arrangements.

If the storage cycle were truly over, we would not be seeing higher prices, growing shipments, a 76% operating margin, and customers signing long-term supply agreements at the same time.

So why did Seagate beat expectations so decisively while SK Hynix disappointed the market?

The most obvious difference is expectations.

Seagate delivered results above expectations and then issued guidance that was also far above expectations. SK Hynix delivered record results, but they still fell short of an already extremely aggressive consensus.

Stocks do not trade simply on whether a company is doing well or badly. They trade on the gap between reality and expectations.

The second difference is supply structure.

The HDD industry is controlled by only a small number of major manufacturers, and capacity expansion has remained disciplined for years. Seagate can also increase storage output through HAMR technology and higher capacity per drive, without building massive new fabrication plants like DRAM and NAND producers must.

As long as cloud customers continue locking in long-term orders, HDD pricing and margins should be easier to defend.

SK Hynix, meanwhile, is preparing to raise its 2026 capital expenditure to nearly KRW 50 trillion, accelerate production at M15X, and expand its Yongin facility from early 2027.

In the short term, this confirms that customers are competing for capacity. In the medium term, however, investors will worry that all this new capacity could eventually bring the storage industry back into oversupply.

Combined with CXMT’s listing and China’s rapid DRAM expansion, even SK Hynix’s leadership in HBM cannot completely protect its traditional DRAM and NAND businesses from concerns that the cycle is approaching its peak.

This is why storage stocks can no longer be treated as one single trade.

Seagate is primarily a bet on the growth of AI-generated data, disciplined HDD supply, and long-term cloud demand. SK Hynix is a bet on HBM leadership, DRAM pricing, and the impact of future capacity expansion. $闪迪(SNDK)$ and Kioxia still depend heavily on whether enterprise SSD demand can absorb future NAND supply.

My view is that Seagate’s results rejected the bearish argument that AI no longer needs storage. SK Hynix’s report rejected the bullish belief that strong fundamentals alone would immediately reverse the sector’s decline.

Storage demand is not dead, but the period when every storage stock could rise together may be over.

From here, the market will focus much more closely on which companies can continue beating expectations, which product categories have the strongest supply discipline, and which businesses are least exposed to Chinese capacity expansion.

The storage industry is not collapsing as one group. Investors are beginning to price each part of the market differently.

Seagate proved that demand is still there. SK Hynix proved that expectations had become too high. Taken together, these two earnings reports show the real situation facing the storage sector today.