The Zhitong Finance App notes that Goldman Sachs made a clear judgment in the newly released outlook report for the US Semiconductor Third Quarterly Report: most segments of the semiconductor ecosystem will see an improvement in profit forecasts, and the risk-reward ratio has improved markedly.
A team of analysts led by James Schneider pointed out that in the past two months, the Philadelphia Semiconductor Index (SOX) fell sharply by 11% during large-scale safe-haven sell-offs, and the S&P 500 index rose 4% during the same period — it was this extreme divergence that completely changed the trading layout before the third quarter from “tactical prudence” to “more constructive” in the second quarter.
Equipment and manufacturing: WFE spending visibility extended to 2028
Goldman Sachs gave semiconductor equipment the clearest direction of upgrading. According to the report, fab equipment (WFE) spending is being locked in early by customers, and industry order visibility has been extended to 2028. Applied Materials (AMAT.US) was listed as one of the top choices, and a target price of $670 was given: Goldman Sachs expects it to raise its long-term profit margin target at SEMICON West on October 13 and chart the WFE market's path to $300 billion. The subsequent mid-November earnings report will provide strong answers based on DRAM/HBM and advanced logic requirements.
Fanlin Group (LRCX.US) ushered in better risk-return after the stock price correction of about 20%. Its GAA surround gate and deposition/etching exposure associated with backside power supply, increased DRAM share, and demand for NAND upgrades formed a triple drive. Goldman Sachs gave it a “buy” rating, with a target price of $380.
In contrast, although KLAC.US (KLAC.US) was expected to exceed expectations for the quarter, it was classified as a tactical failure target due to WFE spending bias towards DRAM and relatively insufficient process control exposure.
Computing power: Agentic AI is revaluing CPUs and ASICs
Under the undertone of cloud vendor capital support for continuous upgrading, Goldman Sachs believes that both server CPUs (benefiting from the increase in Agentic AI penetration rate) and key ASIC projects have room to improve.
AMD (AMD.US) was the most aggressive target in this round. The target price was raised from $640 to $700, maintaining a “buy” rating. Goldman Sachs expects both its third-quarter results and guidance to exceed expectations. The 2027 earnings forecast is $17.10 per share, 7% higher than Wall Street. The core catalyst is the Helios platform's fourth quarter launch and the improvement in server CPU demand and pricing driven by Agentic AI.
Cadence (CDNS.US) is regarded as the target of “least afraid of AI disruption” — Agentic tools are driving the monetization of EDA software. Goldman Sachs expects it to raise the 2026 revenue guide to about 21% year-on-year growth. Goldman Sachs gave it a “buy” rating, and the latest target price is $470.
It is worth noting that although Qualcomm (QCOM.US) may slightly exceed expectations, Goldman Sachs believes that the market's expectations for Agentic AI are too high, and there is a tactical risk of a pullback.
Analog chips: recovery is underestimated, industrial and data centers are dual engines
Goldman Sachs bluntly stated that “Wall Street is still systematically underestimating the recovery slope of analog chips,” and favors companies with the largest industrial, aerospace/defense, and data center exposure. Microchip Technology (MCHP.US) is one of its most popular targets. Industrial inventory replenishment combined with data center revenue impacted the $1 billion target, with a target price of $115.
NXP (NXPI.US) focuses on automotive demand (particularly in the Chinese market) and plans to double data center revenue, with a target price of $325; SiTime (SITM.US) uses MEMS clock replacement logic to combine quartz with Renesas clock business, and expects quarterly revenue to exceed expectations by about 5%, with a target price of $900.
In contrast to this, although Texas Instruments (TXN.US) has strong performance, Goldman Sachs believes that high inventory and high depreciation will drag down its gross margin compared to its peers and maintain sales ratings.
Storage: “money printers” under tight supply, prefer HDD and NAND
In the storage sector, SanDisk (SNDK.US) received a “buy” rating, and the latest target price was as high as $2,200, the highest on the full list: NAND had no new supply recently. Combined with the penetration of enterprise-grade SSDs on the cloud vendor side and the expansion of long-term agreement (LTA) coverage, Goldman Sachs expects its quarterly revenue to exceed expectations by about 6% compared to the midpoint of the guideline.
Seagate (STX.US) received a “buy” rating, with a target price of 960 US dollars. HAMR technology conversion progress is ahead of its peers. Combined with the capital return space after deleveraging, it is expected that it will continue to gain HDD share; in contrast, Western Digital only received a “neutral” rating, and Goldman Sachs clearly “prefers Seagate”.
“Buy” list
Taken together, Goldman Sachs bought a total of 12 rated stocks in this three-quarter reporting season, forming the core basket of its semiconductor configuration:
The computing power and EDA segment includes AMD (AMD.US) and Cadence (CDNS.US);
Analog chips: Microchip (MCHP.US), NXP (NXPI.US), SiTime (SITM.US);
Equipment and materials: Applied Materials (AMAT.US) is included in the Convinced Purchasing List, which also includes Fanlin Group (LRCX.US), Onto Innovation (ONTO.US), Qnity (Q.US), and TER.US (TER.US);
The storage is SanDisk (SNDK.US) and Seagate (STX.US).
At the tactical level, Goldman Sachs is clearly bullish on applied materials, Cadence, and Microchip, which suggests that Qualcomm, Kelei, and Western Digital have downside risks; at the same time, it maintains “sell” ratings for Arm (ARM.US), Texas Instruments (TXN.US), Entegris (ENTG.US), and MKS Instruments (MKSI.US) on the grounds that valuation overdraft, gross margin loss, or growth outperforms the WFE market.
Goldman Sachs's core conclusion is that the two-month deep pullback has washed away crowded long positions, and the fundamentals — WFE spending has moved forward, Agentic AI drives computing power, simulated inventory replenishment, and zero new storage supply — is collectively improving. The three-quarter reporting season will be a window of resonance between bearish recovery and profit improvement.