Morgan Stanley's "Memory Grim Reaper" Shawn Kim Issues Another Warning! The Real Bombshell Is the "2028 Earnings Cliff," Comprehensive Target Price Cuts for Memory Stocks

Morgan Stanley analyst Shawn Kim has once again issued a warning to the memory industry, this time targeting the sector’s “2028 profit cliff.”

According to The Whip Trading Desk, in a newly released report on November 9, Shawn Kim comprehensively lowered the target prices for Samsung Electronics and SK hynix, and upgraded Samsung Electronics to his new top pick, replacing SK hynix. He wrote in the report: "Concerns about memory are less about data center construction delays after the US midterm elections or about AI demand, and more about the risk of the '2028 profit cliff' narrative—the market is skeptical about the sustainability of memory growth momentum because supply is starting to catch up with demand."

He believes that the market debate has shifted from near-term fundamentals to the “2028 profit cliff”—multiple pressures are piling up, including new capacity coming online at the end of 2027, delays in data center construction, and HBM premium narrowing, putting the sustainability of the memory supercycle into question. Based on this, Shawn Kim significantly compressed the long-term profit margin assumptions for SK hynix and Samsung towards historical averages, directly lowering the valuation output of both companies’ models.

This is not Shawn Kim’s first warning to the memory industry during a boom cycle. Wallstreet Insights mentioned that he issued "Memory: Winter is Coming" in August 2021, precisely predicting the subsequent two-year semiconductor downturn; in September 2024, his report on potential HBM oversupply was seen as one of the triggers for the sharp stock declines in Samsung and SK hynix that year; and on July 21 this year, he released another report warning that the AI-driven semiconductor memory frenzy was nearing a turning point, predicting that memory contract prices would peak in the fourth quarter, after which SK hynix and Samsung Electronics shares fell sharply again.

These actions have earned him the nickname "the Grim Reaper of Korean semiconductors" in the Korean market. This report marks yet another warning in the context of continuing de-bubbling of memory-AI trades.

Valuation Reset: Significant Target Price Cuts, Samsung Becomes Top Pick

Shawn Kim made significant adjustments this time. SK hynix’s target price was cut by 38% from 2.6 million KRW to 1.6 million KRW, corresponding to a 2027E P/B of about 1.8x. Compared to the current reference price of 1.132 million KRW, there is still about 41% upside, and the “Overweight” rating is maintained.

The target price for Samsung Electronics common shares was cut by 19% from 381,000 KRW to 309,000 KRW, and the preferred shares’ target price was reduced from 304,800 KRW to 247,200 KRW—both remain “Overweight,” with Samsung Electronics common stock upgraded to the new top pick.

The core driver of the target price cut is Shawn Kim’s significant downward revision to the long-term sustainable profit margin assumptions for both companies. Previously, Shawn Kim assumed a long-term sustainable DRAM profit margin of about 60% before 2035; now, the 2029-2035 average operating margin for DRAM is lowered to around 32% for SK hynix (close to the historical average of 31% for 2010-2025), and to about 39% for Samsung. Meanwhile, the premium of HBM over regular DRAM is assumed to gradually narrow from 20 percentage points in 2029 to zero by 2033.

The report also raises earnings forecasts for both companies for 2026-2028, primarily benefiting from the reduction in share count due to buyback plans. SK hynix’s earnings per share estimates for 2026-2028 change by -1%, +22%, and +29%, respectively; Samsung’s are revised up by 3%, 16%, and 26% for the same years.

“2028 Profit Cliff”: Four Major Pressures Combine

Shawn Kim clearly points out in his report that the “2028 profit cliff” is presently the core debate for memory investments. He lists four key factors that put pressure on long-term profit margin assumptions.

First, supply response is faster than expected. New production capacity will gradually come online around the end of 2027, and recent higher fourth-quarter shipment guidance and faster process migration mean there will be more bits supplied than previously expected. Chinese suppliers entering the consumer market will also put additional pressure on bulk memory pricing.

Second, data center construction delays in 2028. The Morgan Stanley Global Thematic Research team has postponed about 5GW of US data center projects from 2028 to 2029. Accordingly, the global growth rate of newly added data center capacity in 2028 is revised down from about 21% year-on-year to 9%, with a rebound to 20% in 2029. This could weaken the degree of supply tightness and pricing power for memory.

Third, BOM cost constraints limit content growth. Memory now accounts for about 60% of the bill of materials (BOM) cost in AI servers, which is forcing buyers to downgrade from high-density 96GB/128GB configurations to 64GB RDIMM. NAND spot prices have fallen by about 20% in the past two weeks; the 512Gb TLC wafer price is down to about $15, a 27-29% decrease from the March peak, and is lower than the July-August lows, bringing additional pricing risk to the contract market.

Fourth, HBM premium is narrowing. Morgan Stanley has adjusted the HBM operating profit margin forecast for SK hynix from 78-80% down to about 70%, and for Samsung from about 80% down to 65-72%. As competition intensifies and regular DRAM prices normalize, the likelihood of HBM maintaining a high premium continues to decline.

Cycle Positioning: Valuations Have Adjusted, But Earnings Cuts Yet to Come

Despite issuing warnings, Shawn Kim is not entirely bearish. In the report, he provides a more nuanced assessment of the current market position.

The report notes that since the June highs, the Philadelphia Semiconductor Index (SOX) 12-month forward P/E ratio has halved from 26.3x to 13.4x—one of the steepest valuation compressions since the burst of the internet bubble—while consensus forward EPS has actually risen by 24% during the same period. The current forward P/E of about 13x is essentially in line with the median bottom valuations (13.4x) of previous downturns, and is at the lowest level since the Fed’s rate hike low in October 2022.

Valuation has done its job. Now the variable is earnings. Kim reviewed the eleven prior SOX downturns since 2000, finding that nine out of eleven were primarily driven by valuation compression rather than earnings cuts. Only in the 2000-2002 and 2007-2008 collapses did earnings see more than 50% deep cuts due to real demand destruction.

A second leg down would require the 2028 earnings forecast to be cut in half—like in 2000-2002 and 2007-2008—and that is not our base case scenario, but it is why the earnings cut phase rather than valuation is the key area to watch now.

Kim also points out that the breadth of EPS revisions in the AI supply chain has been steadily narrowing since the peak in July 2026: the 3-month average for memory has fallen from 69% to 53%, AI compute from 55% to 45%, networking from 54% to 40%—but all remain positive, that is, earnings forecasts are still being raised, just at a slowing pace.

Samsung vs SK hynix: Late Cycle, Quality Reigns Supreme

In terms of stock selection, Kim has upgraded Samsung Electronics to his top pick, based on a preference in the late cycle for the “largest, most capital-rich, and most scalable technology leader.”

The report states that Samsung is expected to overtake in HBM market share by 2027—according to TrendForce, Samsung’s HBM shipments will grow about 82% to 21 billion Gb (about 39% market share) in 2027, surpassing SK hynix’s 20 billion Gb (about 37%). Meanwhile, Samsung’s more diversified business structure (including a potential recovery in non-memory business) offers additional support. At the current reference price of 188,700 KRW, Morgan Stanley’s 309,000 KRW target price implies about 64% upside.

SK hynix remains “Overweight” as well; Kim believes its HBM leadership, strong 2027 profits, and LTAs (long-term agreements) protecting 2028 still offer support. But the report clearly states that the target price cut “reflects shortened excess profit periods after 2028.”

The report also describes the worst-case scenario: if AI financing costs rise or AI monetization falls short of expectations, buyers may suspend procurement, and already-committed capacity would risk demand shortfalls.