Written by: Rita

In the 7th week of the semiconductor earnings season, two major events are on the horizon: Broadcom and Ambarella will release their quarterly reports sequentially, while Skyworks had already signaled at its technology day last week.

On August 31, Morgan Stanley released its semiconductor weekly report, providing detailed forecasts for the three companies. Broadcom's AI revenue momentum is strong, but market expectations are already at high levels; Ambarella's automotive business continues to grow, with a clear long-term positioning in edge AI; Skyworks has won all RF front-end share for the Google Pixel, with the agreement extending to 2030. Morgan Stanley maintains an overweight rating on Broadcom and Ambarella, and a neutral rating on Skyworks.

Broadcom's AI momentum is strong, and market expectations represent the biggest risk

Morgan Stanley expects Broadcom's AI revenue for the July quarter (F3Q) to be approximately $16 billion, a 48% quarter-over-quarter increase, of which custom ASICs are about $10.8 billion and AI networking is about $5.2 billion. For the October quarter (F4Q), AI revenue is projected to increase another 32% to $21.2 billion, with custom ASICs at about $14.9 billion and AI networking at about $6.2 billion. TPU demand remains strong, and the networking business continues to benefit from the expansion of AI infrastructure.

The real story lies in 2027. Morgan Stanley expects Broadcom's AI revenue in 2027 to be around $120 billion, while the company's previous guidance was "well above $100 billion." Management last quarter hinted that growth would continue through 2028. Morgan Stanley believes that management may not provide an official framework update for 2027 this quarter, but even without it, the fundamental story remains intact. TPU will continue to be the main source of Broadcom's custom ASIC revenue, and Morgan Stanley will closely monitor the progress of several new custom projects, with several expected to ramp up production in 2027.

The biggest risk in the earnings report lies outside of fundamentals; market expectations are the real variable. Last quarter proved a truth: strong performance, but if market expectations are higher, stock prices may not rise. Some investors assume Broadcom's AI revenue will exceed $150 billion in 2027. Morgan Stanley believes this number is possible but not a certainty. Morgan Stanley also reminds not to overinterpret rumors of downward revisions to CoWoS capacity forecasts; these supply chain data have significant noise, and Broadcom's growth momentum remains very strong.

TPU share is another ongoing controversy. MediaTek's participation is real, and rumors of AMD's involvement in TPU v10 and Marvell's warrant agreement with Google suggest that Google is expanding its supplier ecosystem. However, Morgan Stanley still expects Broadcom to retain about 80% of the long-term TPU opportunities. There is no substantive risk to its core TPU share. Morgan Stanley lists Broadcom as the second preferred choice in the AI computing space, second only to NVIDIA.

Morgan Stanley's specific forecast for Broadcom's July quarter: revenue of $29.4 billion, roughly in line with the market consensus of $29.2 billion; gross margin of 74.0%, slightly above the market consensus of 73.5%; earnings per share of $3.24, slightly above the market consensus of $3.22. October quarter forecast: revenue of $34.8 billion, slightly above the market consensus of $34.7 billion; gross margin of 72.9% in line with market consensus; earnings per share of $3.85, slightly above the market consensus of $3.84.

Ambarella's automotive growth continues, with a clear long-term layout in edge AI

Morgan Stanley expects Ambarella's revenue for the July quarter (F2Q) to grow 7.9% quarter-over-quarter, with an additional 12% increase in the October quarter. Automotive is the main growth engine this year, with revenue growth of approximately 20%, mainly driven by fleet telematics, AI video and safety applications, and an increase in AI content per vehicle.

The Internet of Things trend is somewhat polarized. Enterprise security is relatively healthy, while the consumer market may face pressures from rising memory and storage costs in the second half of the year. Morgan Stanley believes this is the main short-term risk, although the extent of the impact is still uncertain.

Long-term, Morgan Stanley believes Ambarella is well-positioned in its years-long expansion in edge AI. Drones are the most attractive short-term opportunity, with growth occurring in both consumer and commercial sectors; the upcoming ramp-up of CV7 will provide new growth sources and support an upward trend in ASP. Automotive remains an important long-term opportunity, but there were no significant updates on the large-scale ramp-up of CV3 autonomous driving last quarter. Overall, Ambarella's story is transitioning from a high-growth phase of the product cycle to a stable execution phase, but Morgan Stanley remains optimistic about its technology positioning in low-power edge AI inference.

Regarding recent unconfirmed merger rumors, Morgan Stanley believes that Ambarella's strategic value has always been part of its overweight rating, although confidence mainly stems from technical strength. While the veracity of the rumors cannot be validated, Morgan Stanley believes there is potential rationale for such a combination. Ambarella offers differentiated edge AI and computer vision capabilities, while NXP's scale, established automotive relationships, and OEM trust can accelerate commercialization, especially in the automotive field where Ambarella has historically struggled to compete with large suppliers.

Morgan Stanley's specific forecast for Ambarella's July quarter: revenue of $108.3 million, exceeding the market consensus of $107.8 million; gross margin of 59.8% in line with market consensus; earnings per share of $0.16, slightly below the market consensus of $0.17. October quarter forecast: revenue of $121.2 million, exceeding the market consensus of $119.3 million; gross margin of 59.7% essentially in line with market consensus of 59.6%; earnings per share of $0.29, slightly below the market consensus of $0.30.

Skyworks wins all RF front-end share for Google Pixel

Morgan Stanley attended Skyworks' technology day in San Jose last week. The company focused on showcasing its technology portfolio across the mobile and broad markets.

On the mobile side, the most noteworthy achievement is Skyworks' overall victory with Google Pixel. For the first time, Skyworks has won the mid-band position during Google's approximately two-year re-bidding cycle, currently holding nearly all RF front-end share for Pixel, with the agreement extended to 2030, and expected cumulative revenue to exceed $1 billion. The company also emphasized its close collaboration with MediaTek and early visibility into next-generation RF demand.

Beyond mobile, the presentation focused on how Skyworks applies RF, timing, isolation, sensing, and power capabilities to faster-growing markets. The precision timing area focuses on single-chip integrated network synchronization, jitter attenuation, clock generation, and software; data centers focus on higher voltage power architectures covering isolation, gate driving, and sensing, with Skyworks positioned in all aspects up to 800V reaching from the grid to rack stacking. The company also discussed opportunities in automotive, WiFi/broadband, drones, and other edge connectivity applications.

Morgan Stanley holds a positive view on Skyworks' long-term opportunities, but the broad market remains polarized recently, with strength in data centers and automotive, while IoT is relatively weak; the smartphone environment maintains resilience under memory pressure. In the long run, the broad market is expected to become a larger business segment for Skyworks, helping the company reduce reliance on the mobile side, especially given Apple's continued loss of share. The Qorvo transaction and integration remain key catalysts. Morgan Stanley believes patient investors will be able to see value, but the remaining business pressures lead it to maintain a neutral rating.

Disclaimer

This article is a compilation and interpretation of third-party brokerage research reports (Morgan Stanley, August 31, 2026) by潮向研究, combining information from the public market. The ratings, target prices, earnings forecasts, and related judgments cited in the article reflect the views of the brokerage's analysts, represent the stance of their respective institutions, do not represent the views of潮向研究, and do not constitute any investment advice.

The market carries risks; decisions should be independent. This article should not be used as the basis for buying or selling any securities.