Goldman Sachs is bullish that the semiconductor rebound will continue: why are memory chips and equipment stocks the top choice for funds?

On July 23, Goldman Sachs explicitly said that the current semiconductor rebound rally is likely to continue. The backdrop to this call is that since mid-June, hedge funds have given back about 80% of their cumulative net buying volume in global semiconductor and semiconductor equipment stocks since the beginning of the year. The previously extremely crowded positioning has been clearly cleaned up. In the past one or two days, however, funds have started buying again, and Goldman Sachs believes this demand may continue.

The most concentrated buying direction is two major themes that previously saw the sharpest pullbacks—memory chips and semiconductor equipment. The most notable buy activity recently has been in memory stocks such as Seagate Technology (STX), Western Digital (WDC), Micron Technology (MU), and SanDisk (SNDK), as well as equipment stocks including Applied Materials (AMAT), ASML, and Lam Research (LRCX).

This flow of funds echoes the recent market trend. After AI and chip stocks faced continuous pressure, memory stocks such as Micron, SanDisk, and Seagate rebounded sharply at one point, and the Philadelphia Semiconductor Index (PHLX Semiconductor Index, SOX) also recorded a strong recovery. After the semiconductor sector went through a round of “violent deleveraging,” why did capital choose memory and equipment as the priority areas to refill? Which key variables determine the sustainability of this rebound? This article will systematically break down these questions by combining Goldman Sachs’ latest positioning data, the memory price trend, and the logic behind AI capital expenditures.

Semiconductor positioning: extreme crowding has eased, and fast deleveraging is already mostly done

To understand the logic behind the current return of funds, the first step is to clarify the root cause of the earlier semiconductor sector correction.

In the first half of 2026, AI chip stocks kept rising, and capital became highly concentrated in the semiconductor sector. Goldman Sachs data shows that the net allocation to global semiconductor and semiconductor equipment stocks as a share of global Prime Book rose to about 10% at the start of the year, then climbed to a historical high of about 24% in June. In the U.S., net allocation to semiconductor and equipment stocks rose from about 7% at the start of the year to about 14% in June.

This extreme concentration means that once market sentiment shifts at the margin, the pressure for funds to take profits will be released in a concentrated manner. The semiconductor sector then went through a brutal deleveraging cycle—large-scale selling by hedge funds, with the previously crowded long positions being heavily cleaned out.

As of July 23, the net allocation to global semiconductor and equipment stocks had fallen from the June historical high to about 19%, but it remains at the 84th percentile over a one-year period and the 97th percentile over a five-year period. Net allocation for U.S. semiconductor and equipment stocks has fallen to about 11%, corresponding to the 79th percentile over a one-year period and the 96th percentile over a five-year period.

Goldman Sachs believes the signal from this set of data is: semiconductor positioning has not yet been brought down to a “cheap” range, but fast deleveraging has already been done for the most part. This means that as long as subsequent earnings reports and AI capex guidance do not continue to worsen, some capital will likely prioritize refilling the sections that were hit hardest, but still have fundamentals support.

From a more macro perspective, Goldman Sachs previously made it clear that investors are underweighting U.S. tech stocks—especially the “Seven Tech Giants”—and instead favor semiconductor companies that can benefit from AI investment. This long-term allocation logic has not changed due to the short-term adjustment.

Memory chips: “inelastic demand” from AI data center expansion

Memory chips have become the first mainline for capital to refilling, supported by clear industry logic.

As AI data center compute capacity expands, it requires not only compute chips such as GPUs, but also matching storage capabilities. With AI model sizes continuing to expand, compute demand drives an increase in the number of GPUs, and the increase in GPUs in turn boosts the accompanying demand for high-bandwidth memory (HBM), DRAM, and NAND flash.

This logic is translating into tangible performance. The memory industry has strong cyclicality, and the core issue the market is actively debating right now is whether AI demand will permanently change the industry demand curve. Based on current data, the supply-demand balance for memory chips remains tight. In the second quarter of 2026, LPDDR4 and LPDDR5 prices were up about 1x cumulatively versus the fourth quarter of 2025, with the Q2 single-quarter increase of roughly 58% to 63%. As of July 21, the DDR5 16G (2Gx8) 4800/5600 price first broke above the $50 mark, reaching $50.07; the DDR4 16Gb (2Gx8) 3200 price rose to $81.20—both setting historical highs.

Constraints also exist on the supply side. SK hynix Chairman Choi Tae-won recently said that next year, AI semiconductor demand is expected to grow year over year by 60% to 100%, and overall memory semiconductor demand growth will also reach 50% to 60%. He also warned that the AI-driven semiconductor supply-demand gap will further widen next year, putting additional upward pressure on memory prices.

Citi Securities’ key components chief analyst pointed out that memory chip prices have surged, and demand from AI data centers continues to squeeze capacity for advanced flash and memory, meaning the industry’s supply-demand imbalance—where supply falls short of demand—will likely persist at least until 2028. Morgan Stanley also holds a similar view, saying that tightness in the memory chips’ supply-demand situation will last until 2028.

From market performance, the rebound strength in memory stocks has been the most pronounced. On July 21 Beijing time, U.S. memory stocks kept climbing: Micron Technology rose 10.17%, SK hynix rose 11.18%, SanDisk rose 11.86%, Western Digital rose 12.67%, and Seagate Technology rose 10.42%. Goldman Sachs also raised target prices for several memory giants, lifting SanDisk’s target price from $1,200 to $2,200, and Western Digital’s target price from $400 to $650.

Semiconductor equipment: the “water-seller” logic of foundry capacity expansion

Semiconductor equipment has become the second mainline for capital to refill, with equally clear logic: as chip manufacturing expands, foundry investment increases, and equipment orders grow accordingly.

The surge in AI chip demand is driving global foundries into a new round of expansion. Whether it’s logic chips, memory chips, or power semiconductors, all capacity expansion is inseparable from core semiconductor manufacturing equipment such as lithography tools, etchers, and thin-film deposition equipment. ASML, the world’s only supplier of extreme ultraviolet (EUV) lithography systems, is in an irreplaceable position in this expansion cycle.

Among the most notable buys that Goldman Sachs’ trading desk has observed recently, equipment stocks occupy an important place—Applied Materials, ASML, and Lam Research all appear on the list. These three companies cover different key stages of semiconductor manufacturing: Applied Materials focuses on materials engineering and deposition equipment; Lam Research targets etching and deposition; and ASML dominates the advanced lithography market.

From a fundamental perspective, equipment stock performance still has support. ASML’s net revenue in the second quarter of 2026 reached €2.92 billion, reflecting strong profitability. Applied Materials’ share price has risen 187.59% over the past 52 weeks. Analysts gave Applied Materials a “Strong Buy” rating, with an average target price of $623.06, implying about 12.5% upside from the current price.

Goldman Sachs previously raised Applied Materials’ target price from $520 to $645. This raise itself reflects the optimistic view of sell-side institutions on the outlook for semiconductor equipment demand—because the foundry expansion cycle has not ended, and the trend of equipment order growth is expected to continue.

Can the rebound continue? Three key observation variables

In its quick commentary, Goldman Sachs clearly pointed out three key dimensions to watch to determine whether the semiconductor rebound rally can continue.

First, whether cloud providers continue to raise their AI capital expenditure plans. The capex plans of Alphabet, Meta, Microsoft, and Amazon are the “ceiling” indicator for semiconductor demand. The latest data shows that Alphabet has raised its 2026 capex guidance from $180 billion–$190 billion to $195 billion–$205 billion, with the midpoint rising to $200 billion, and expects 2027 capex to grow significantly as well. Combined, the four cloud giants’ 2026 capital spending could surpass $725 billion. As long as this upward trend is not broken, downstream demand for semiconductors will have basic support.

Second, whether memory prices and orders can support profit expectations. The rise in memory stocks such as Micron and SanDisk ultimately needs to show up at the earnings level. Memory chip prices are still in an upward channel, but what the market needs to see is the persistence of orders and the stability of pricing. If memory prices loosen, or order growth slows, valuations of memory stocks will face renewed pressure to be reassessed.

Third, whether equipment stock orders can validate that AI data center expansion is still progressing. Equipment stock performance is the most direct “evidence” of foundry capacity expansion. If order data and earnings guidance from ASML, Applied Materials, and others continue to beat expectations, it will validate that AI data center expansion has not slowed down, providing fundamental support for the entire semiconductor sector.

Conclusion

After a round of severe adjustments in the semiconductor sector—from extreme crowding to rapid deleveraging—hedge funds have already given back about 80% of their net buying volume since the start of the year. While positioning has not yet returned to a “cheap” range, extreme crowding has clearly eased. Under this backdrop, funds have started buying again in the two mainlines that previously saw the sharpest pullbacks: memory chips and semiconductor equipment.

Memory chips benefit from the inelastic demand for HBM, DRAM, and NAND from AI data centers, alongside capacity constraints on the supply side—so prices remain in an upward channel. Semiconductor equipment benefits from the continued progression of the foundry expansion cycle, with a clear trend of rising equipment orders.

But the sustainability of the rebound is not without doubt. Whether cloud providers continue to raise AI capex, whether memory prices and orders can support profit expectations, and whether equipment stock orders can confirm that AI data center expansion is still moving forward—these three variables will determine whether this semiconductor rebound is a brief technical repair or a trend-based valuation reshaping.

On July 23 Beijing time, in the crypto market, Bitcoin was $66,188.99, down 0.32% over 24 hours; Ethereum was $1,947.90, up 1.31%. Global crypto total market capitalization is about $2.33 trillion. While the semiconductor sector rebound and crypto market volatility are different asset categories, both reflect the process of repricing global risk assets amid macro uncertainty. For investors, understanding how capital flows across different sectors is more valuable in the long run than chasing short-term price swings.

FAQ

Q: Why have semiconductor stocks rebounded recently?

AI chip stocks continued rising earlier, causing semiconductor positioning to become highly concentrated; then profit-taking by institutional funds triggered the correction. As hedge funds gave back about 80% of their net buying volume, extreme crowding has clearly eased. Goldman Sachs believes that rapid deleveraging has already been completed for the most part, and some funds have started refilling areas with fundamentals support even if they were hit hardest.

Q: Why are funds prioritizing buying memory chip stocks?

AI data center expansion not only needs GPUs, but also matching storage chips such as HBM, DRAM, and NAND. Memory chip prices are still in an upward channel, and prices for both DDR5 and DDR4 chips have set historical highs. Supply-demand gaps are expected to persist at least until 2028.

Q: What is the investment logic for semiconductor equipment stocks?

Chip manufacturing expansion drives increased foundry investment, which then boosts equipment order growth. Equipment leaders such as ASML, Applied Materials, and Lam Research are in core beneficiary positions in this expansion cycle. As long as the AI data center expansion trend does not change, the demand fundamentals for equipment stocks will have support.

Q: Can the semiconductor rebound rally continue?

Goldman Sachs highlights three key observation variables: whether cloud providers (Alphabet, Meta, Microsoft, Amazon) keep raising AI capex; whether memory prices and orders can support profit expectations; and whether equipment stock orders can validate that AI data center expansion is still underway. All three are necessary.

Q: What is the supply-demand situation for memory chips?

Current memory chip supply is tight, and ongoing AI data center demand continues to squeeze capacity for advanced flash and memory. SK hynix expects next year’s AI semiconductor demand to grow year over year by 60% to 100%. TrendForce expects DRAM to rise 13% to 18% quarter over quarter in the third quarter, and NAND to rise 10% to 15%.