Morgan Stanley: AI Ignites a "New Product Cycle" for Smartphones

It's time to re-evaluate investment opportunities in smartphones!

Morgan Stanley believes that Edge AI will be the next popular trend after cloud intelligence, igniting a "new product cycle" for smartphones.

Edge AI refers to the execution of artificial intelligence tasks on the edge of devices or systems (such as smartphones and personal computers), rather than relying solely on data centers. Currently, most AI functions can only be used in the cloud.

According to a research report released by Morgan Stanley analyst Andy Meng's team last week, Edge AI technology is expected to be launched next year, bringing professional-level photography, desktop gaming, and other new applications to smartphones. It is estimated that Edge AI will start leading the next product cycle of smartphones from 2025.

Coupled with the continued slowdown in inflation boosting demand, destocking is expected to be completed in the third quarter of this year. Morgan Stanley predicts that smartphones will recover from 2024, and the institution has also raised its smartphone shipment forecast for the next two years.

Specifically, Morgan Stanley predicts that global smartphone shipments will recover to 3.9% in 2024 and 4.4% in 2025, which is more optimistic than the overall market forecast.

In addition, Morgan Stanley remains bullish on Apple, believing that Apple's reliance on the Apple product line and the leading position of the M-series chips will make it the biggest winner in the field of Edge AI.

Why Edge AI?

Currently, AI applications typically require the use of the cloud, and Edge AI has not been widely used in large language models (LLMs). According to Morgan Stanley's research report, compared with cloud intelligence, Edge AI has four key advantages:

Reduced or eliminated cloud costs: Edge AI can be executed locally on the device, reducing reliance on cloud resources and thus reducing or eliminating costs associated with cloud computing. This makes fine-tuning AI models and application development more cost-effective.

Reduced latency: Generating responses to queries on the device instead of waiting for remote data centers to generate results can significantly reduce communication latency.

Enhanced privacy protection: Keeping processing on the device when using personal data to improve generative responses helps enhance privacy.

Balancing personalized expression and privacy rights: Edge AI allows applications to meet users' personalized expression and characteristics without sacrificing privacy.

Morgan Stanley believes that Llama2 is likely to become the Android of the mobile era, as it is the most cost-effective open-source LLM to date, with performance roughly equivalent to GPT 3.5. This means that it can provide advanced natural language processing capabilities at a relatively low cost. Especially, Llama2 has entered into a strategic partnership with Qualcomm. Qualcomm holds a leading position in the non-Apple smartphone market and its Snapdragon platform, supported by the Hexagon processor, has powerful artificial intelligence capabilities. This collaboration creates a synergistic effect, building an ecosystem for the development of large-scale smartphone applications and enhanced features.

A Glimpse into the Future of AI Smartphones

By embedding edge artificial intelligence, smartphones of the future will achieve more enhanced functionalities and new features, playing a role more like a virtual companion. They will not only be able to understand and meet users' needs but also provide suggestions in areas such as photography, language translation, and navigation.

Some of these features include:

Professional-grade Photography: Utilizing AI-driven cognitive ISP (Image Signal Processing) technology, real-time semantic segmentation is achieved, optimizing the quality of photos and videos, allowing for fine adjustments to every detail.

Desktop-level Mobile Gaming: Smartphones can provide a desktop-level gaming experience, including real-time ray tracing technology, multiple dynamic shadows, and desktop-level motion blur, to create hyper-realistic graphics and provide smoother interaction, outstanding performance, and enhanced audio.

Always-aware Camera: New features can automatically activate or lock the screen by detecting whether the user is looking at the device. It can also automatically lock the screen when the user is not nearby, protecting user privacy. In addition, it supports gesture control in situations where the user cannot hold the device and automatically launches relevant applications when triggered by barcode or other data detection.

Stable Diffusion: This is an AI model for text-to-image generation that allows smartphones to create realistic images based on any text input in a matter of seconds. It can be used for image editing, retouching, style transfer, super-resolution, and other applications. However, due to the need for over 1 billion parameters, stable diffusion is primarily limited to running in the cloud. In February 2023, Qualcomm made stable diffusion possible on Android phones for the first time through Snapdragon's full-stack AI optimization. Voice Assistant Becomes Smarter: Qualcomm's AI engine and hardware-accelerated audio capabilities can help create more intelligent, responsive, and personalized voice assistants. It can not only recognize and learn users' speech and intonation patterns but also automatically launch appropriate applications through context audio detection, providing a more intelligent voice interaction experience.

With the application of edge artificial intelligence, smartphones are expected to become more personalized and intelligent tools to meet users' daily life and creative needs. Morgan Stanley believes that the application of artificial intelligence will lead the next product cycle of smartphones starting from 2025, which will be a key catalyst for stock revaluation.

Triggered by Edge Artificial Intelligence, Will a New Round of Recovery Begin?

After reaching its peak in 2016, the smartphone industry experienced five consecutive years of declining shipments, followed by a one-year recovery in 2021 (due to the COVID-19 pandemic). Looking ahead, Morgan Stanley believes that the completion of destocking and temporary component orders has laid a solid foundation for recovery, and edge artificial intelligence is likely to be a key triggering factor.

Considering the cyclical demand recovery brought by downward inflation and the positive impact of edge artificial intelligence, Morgan Stanley has raised its global smartphone shipment forecast.

We currently expect smartphone shipments to recover by 3.9% in 2024 and 4.4% in 2025.

In terms of regional breakdown, we expect developing markets such as India, the Middle East, and Africa to be the main growth drivers, while developed markets will have lower growth rates due to their already high smartphone penetration.

Component Suppliers vs. Phone Manufacturers, Android vs. Apple

Due to a wave of intense stock price adjustments, smartphone component suppliers may outperform smartphone manufacturers in the latest recovery cycle.

Since the decline of the smartphone market in 2022, suppliers have experienced the largest stock price adjustments. For example, from January 2022 to September 2023, the stock prices of Qualcomm, MediaTek, Sunny, Qtech, and Wingtech all experienced significant declines, while the stock prices of smartphone manufacturers performed relatively stronger. The weakness in the stock prices of smartphone suppliers indicates a negative market sentiment towards these companies and the industry as a whole, with low expectations. However, when the smartphone industry recovers, they are likely to be the major beneficiaries. Among the many stocks related to the supply chain, companies that are involved in edge artificial intelligence products may outperform other stocks.

Considering the greater data processing capabilities required for LLM in artificial intelligence, Morgan Stanley expects memory manufacturers to be the major beneficiaries, with the upgrade of RAM (Random Access Memory) specifications becoming an investment theme worth paying attention to.

At the same time, Morgan Stanley also believes that the Android supply chain is more likely to outperform the Apple supply chain.

Android shipments experienced a sharp decline in 2022 (more than 10% decline), while Apple showed more resilience (4% decline). Once the recovery begins, we believe that the Android supply chain is more likely to achieve stronger sales, average selling prices, and profit margin recovery, which could significantly increase profitability. At the same time, valuations may be re-rated, which could greatly boost stock performance. The Apple supply chain will also be positively affected, but not as significantly as Android.

We also expect Huawei's return to the market to help the Android supply chain outperform the broader market. According to our latest AlphaWise survey, approximately 2,000 Chinese consumers are willing to switch phones in the next 12 months (NTM), with the proportion of all survey respondents increasing from 43% in 2022 to 46% in 2023. Among all domestic manufacturers, Huawei is the most popular brand.

Although Morgan Stanley is not optimistic about the Apple supply chain, the institution still believes that Apple itself will become a leader in edge artificial intelligence, based on the leading position of its product series and M-series chips. Morgan Stanley's Erik Wooding has given Apple a buy rating:

As LLM scales down and moves to the edge to improve latency and inference costs, smartphone performance and battery life will become key differentiating factors, as LLM requires enhanced computing, more memory, and quickly drains device batteries. We believe that the iPhone may gain more market share due to Apple's in-house chip development and optimization of its hardware for artificial intelligence, resulting in better performance and longer battery life.

As artificial intelligence applications move towards edge artificial intelligence, we believe this is a key advantage. We believe that Apple can monetize the new generation of artificial intelligence through hardware share revenue, new ways of acquiring traffic acquisition costs (TAC), better service monetization, App Store purchases, and premium Siri subscription services.

Morgan Stanley estimates that for every 1 percentage point increase in Apple's smartphone market share, it will bring in 12 million units of iPhone sales, which is equivalent to $11 billion in revenue, or earnings per share of $0.13 to $0.19.