After a two-month drop of 11%, Goldman Sachs reverses stance and goes long: Is a "golden dip" coming for the semiconductor sector?
Goldman Sachs has turned positive on the Philadelphia Semiconductor Index after an 11% decline over the past two months, shifting its third-quarter outlook. The bank is optimistic about equipment, storage, and analog sectors, issuing "Buy" ratings for 12 stocks including AMD, while remaining bearish on Arm and Texas Instruments.
Odaily Finance APP has noted that Goldman Sachs made a clear judgment in its latest US semiconductor Q3 earnings preview: most sub-sectors in the semiconductor ecosystem are poised for upward earnings revisions, and the risk-reward profile has significantly improved.
The analyst team led by James Schneider pointed out that over the past two months, the Philadelphia Semiconductor Index (SOX) plunged 11% amid broad risk-off selling, while the S&P 500 rose 4%—it’s this drastic divergence that has turned pre-earnings positioning from “tactically cautious” in Q2 to “more constructive” now.
Equipment & Manufacturing: WFE Spending Visibility Extends to 2028
Goldman Sachs gives its clearest upgrade direction to semiconductor equipment. The report notes that wafer fab equipment (WFE) spending is being locked in by customers ahead of time, and industry order visibility has extended to 2028. Applied Materials (AMAT.US) is listed as one of the top picks, with a target price of $670: Goldman expects the company to raise its long-term margin targets at SEMICON West on October 13, and outline a path for the WFE market toward $300 billion. The following mid-November results are expected to deliver strong performance on the back of DRAM/HBM and advanced logic demand.
Lam Research (LRCX.US) sees an improved risk-reward after a pullback of about 20% in its stock price. Its exposure to GAA (Gate-All-Around) and backside power, increased DRAM share, and NAND upgrade demand offer a threefold catalyst. Goldman assigns it a “Buy” rating with a $380 target price.
In contrast, KLA (KLAC.US) is expected to beat earnings expectations, but with WFE spending biased toward DRAM and relatively limited process control exposure, it is listed as a tactical underperformer.
Compute: Agentic AI is Re-evaluating the Value of CPU and ASIC
Against the backdrop of continued upward revisions to cloud vendors’ capital expenditures, Goldman believes there’s upside for server CPUs (benefiting from rising Agentic AI penetration) and key ASIC projects.
AMD (AMD.US) is the biggest upgrade this round, with its target price raised from $640 to $700 and a maintained "Buy" rating. Goldman expects both Q3 results and guidance to beat expectations, with its 2027 EPS forecast ($17.10) 7% higher than Wall Street consensus. The main catalysts include ramping of its Helios platform in Q4, and demand and pricing improvement in server CPUs driven by Agentic AI.
Cadence (CDNS.US) is seen as the “least afraid of AI disruption”—Agentic tools are driving monetization of EDA software. Goldman expects it to raise its 2026 revenue guidance to around 21% YoY growth. Goldman assigns it a "Buy" rating with a latest target price of $470.
It is worth noting that while Qualcomm (QCOM.US) may modestly beat expectations, Goldman believes market expectations for Agentic AI are too high, suggesting tactical downside risk.
Analog Chips: Recovery Underestimated, Dual Engines from Industrial and Data Center
Goldman directly states that “Wall Street is still systematically underestimating the slope of analog chip recovery,” and prefers companies most exposed to industrial, aerospace/defense, and data center segments. Microchip Technology (MCHP.US) is one of its top picks. Industrial restocking plus data center revenue impact targets $1 billion, with a target price of $115.
NXP (NXPI.US) focuses on automotive demand (especially in China) and plans to double data center revenue, with a target price of $325. SiTime (SITM.US), leveraging MEMS clocks as a quartz replacement and the consolidation of Renesas' clock business, is expected to beat quarterly revenue by about 5%, with a target price of $900.
In contrast, while Texas Instruments (TXN.US) has strong performance, Goldman believes high inventory and depreciation will drag its gross margin under peers, maintaining its "Sell" rating.
Memory: “Money Printing Machine” Under Tight Supply, Prefers HDD & NAND
In memory, SanDisk (SNDK.US) is rated “Buy” with a latest target price as high as $2,200, the highest in the list: NAND has seen no new supply recently, and with enterprise SSD penetration among cloud vendors and broader long-term agreements (LTA), Goldman expects quarterly revenue to exceed guidance midpoint by about 6%.
Seagate (STX.US) receives a “Buy” rating, target price $960. Its HAMR technology transition leads the industry, and post-deleveraging capital return could help further gain HDD share. By contrast, Western Digital is only given a “Neutral” rating, with Goldman clearly stating a “preference for Seagate.”
“Buy” List
In summary, Goldman gives “Buy” ratings to 12 stocks this Q3 earnings season as the core basket for semiconductor allocation:
Computing & EDA: AMD (AMD.US) and Cadence (CDNS.US);
Analog chips: Microchip Technology (MCHP.US), NXP (NXPI.US), SiTime (SITM.US);
Equipment & Materials: Applied Materials (AMAT.US) as a conviction buy, along with Lam Research (LRCX.US), Onto Innovation (ONTO.US), Qnity (Q.US), and Teradyne (TER.US);
Memory: SanDisk (SNDK.US), Seagate (STX.US).
On a tactical level, Goldman is overtly bullish on Applied Materials, Cadence, and Microchip Technology, and flags downside risks for Qualcomm, KLA, and Western Digital; it also maintains “Sell” ratings on Arm (ARM.US), Texas Instruments (TXN.US), Entegris (ENTG.US), and MKS Instruments (MKSI.US), citing overvalued multiples, lagging gross margins, or growth trailing the WFE sector.
Goldman’s core conclusion is that two months of deep correction have eliminated crowded long positions, while fundamentals—advanced WFE spending, Agentic AI boosting compute power, analog restocking, and zero new supply in memory—are all trending up. The Q3 earnings season will be a window for short covering and synchronized upward revisions of earnings estimates.
Disclaimer: The content of this article solely reflects the author's opinion and does not represent the platform in any capacity. This article is not intended to serve as a reference for making investment decisions.
You may also like
XRP News Today : Ripple Expands Prime Brokerage Services for Brevan Howard
ROI - For the Trump-led Treasury, the "tail" of the auction is the most difficult part: McKeever
The views expressed in this article are solely those of the author, Reuters columnist Jamie McGeever. Reuters, Orlando, Florida, October 6 – U.S. Treasury auctions are typically dull, predictable, and not newsworthy. But these are not ordinary times, and the Trump administration now faces the risk of sluggish U.S. debt sales making headlines. The U.S. Treasury plans to issue nearly $120 billion in Treasuries this week, the first non-bill bond sales in two weeks: $58 billion in three-year notes on Tuesday, $39 billion in ten-year notes on Wednesday, and $22 billion in thirty-year bonds on Thursday. These auctions would usually be insignificant events, but due to the exceptionally weak performance of auctions from September 22 to 24—especially the five-year note auction on September 23, which triggered the largest spike in bond yields since April last year—they are attracting growing attention. Since then, yields have not only failed to retreat but have surged across most tenors to multi-decade highs. It's worth noting that the possibility of a U.S. Treasury auction "failing" is almost zero. Primary dealers—currently 26 Wall Street banks and institutions authorized by the New York Fed as market makers for Treasuries—are always involved. They effectively underwrite the sales, ensuring the smooth operation of the $30 trillion U.S. Treasury market, the most liquid market in the world. This, in turn, allows the entire global financial system to function, given that trillions of dollars in global debt, assets, and market derivatives are benchmarked against U.S. Treasuries. Treasuries are also the primary collateral for lubricating the financial “pipes” of the U.S. and global markets, including repo agreements, interbank loans, and financing. In short, as long as U.S. Treasuries remain the pillar of the global financial system, there will always be buyers at Treasury auctions. The question, as always, is at what price these bonds will be sold. Currently, borrowing costs in the secondary market are at their highest levels since the mid-2000s, so it's reasonable to expect that the Treasury will pay relatively high rates in the primary market as well. But as recent auctions have shown, negative surprises remain possible. "Too big to be absorbed by the market"? The $70 billion five-year auction on September 23 was among the most worrisome in years. Demand, as measured by the bid-to-cover ratio, was at a nine-year low. The Treasury ended up selling the notes at a yield of 5.033%, more than 3 basis points above the market yield at the close of bidding. Three basis points might not sound like much, but it's exceptional for a five-year note auction. This is the largest so-called "tail" since June 2022. According to JPMorgan analysts, the last time a five-year auction had a three-basis-point tail was back in 2011—amid the brewing debt ceiling crisis that eventually led to a U.S. credit rating downgrade in August that year. Currently, concerns over the U.S.'s daunting fiscal outlook are driving up long-term borrowing costs. As a result, markets generally expect the Trump administration to gradually shift the Treasury’s massive funding needs toward the lower-yield (and therefore lower-cost) short- and medium-term segments of the curve. That's why the five-year note auction two weeks ago sparked such concern. A three-basis-point tail is common in long bond auctions, but not in the "belly" of the yield curve. If the Treasury is forced to pay a higher premium to issue these bonds, then Houston, we have a problem. A large auction tail can be caused by many factors, including market volatility on the day of the auction or more concerning, fundamental issues that may erode demand over time. The two are often hard to distinguish because they are not mutually exclusive. On a brighter note, this unease has not yet spread to the short end of the yield curve. At least, not yet. Three-year and ten-year Treasury yields are up about 50 basis points from the last auction a month ago, hovering around 4.96% and 5.32%, respectively. The thirty-year yield is up roughly 35 basis points to 5.65%. These levels should be high enough to attract strong demand and ensure smooth sales, right? Maybe. But if surprises do occur, volatility and uncertainty could spill over across the market. Investors will be watching developments as closely as hawks. (The views in this article are solely those of the author, a Reuters columnist.) Like this column? Check out Reuters' "Unhedged" (ROI), your essential new source for global finance commentary. Follow ROI on LinkedIn and X. You can also listen to the daily "Morning Bid" podcast on Apple, Spotify, or the Reuters app—subscribe for in-depth market and finance news, seven days a week. US 5-year auction has biggest 'tail' since 2022 https://fingfx.thomsonreuters.com/gfx/mkt/dwpkmkzogpm/TAIL.png (For the convenience of non-English speakers, Reuters provides automated translations of its reports
Crypto prices
More
