AI-Driven Apple in the Replacement Cycle

During the surge in NVIDIA's stock price after the stock split, the companies supplying NVIDIA also reached historic highs. These include TSMC in the manufacturing sector, AMD and DELL in the server segment, and only three companies capable of supplying HBM - SK Hynix, Samsung, Micron, and ARM in the architecture segment.

After various tech giants in the US stock market increased their capital expenditures, these companies have been making a lot of money, at least in the short term, with no worries about orders. The hardware arms race continues to accelerate. However, the development on the application side is not as fast as the hardware side. Apart from the successful commercialization of ChatGPT, most AI applications still have weak profitability. There has yet to be a company that can provide a large number of users with an entry point on the application side. The market currently expects Apple, which entered the AI field latest, to excel in the application side.

I. Apple ALL IN AI, Game-changing Cycle of Phone Upgrades

Two weeks ago, Apple launched the Apple Intelligence program related to AI. This technology refers to embedding large AI models in iPhones starting from the iPhone 16 series. Part of the technology comes from Apple's own research and development, namely the original SIRI, while another part requires partners to provide jointly.

After announcing the AI program, Apple immediately attracted large models such as ChatGPT, Google's Claude, Meta's Llama, and Perplexity to apply for cooperation with Apple. Apple's stock price rose by over 10% in two days, reaching a new all-time high.

The main reason for the surge is Apple's accelerated implementation of the AI program. The market highly recognizes Apple's ability to produce excellent products, coupled with the overall rise in tech stocks, making the AI program a catalyst.

Although Apple is the latest tech giant to fully invest in AI, Apple has the user base with the highest APRU in the smartphone industry. Users with strong payment capabilities are exactly what large model companies need. Therefore, Apple has considerable room for imagination in terms of payment on the AI application side. According to forecasts, if Apple's AI features implemented at the conference are integrated into new models, it is expected to drive a 5-10% increase in iPhone sales.

The combination of these two points has led foreign investors to speculate that after embedding large AI models, Apple can drive more consumers to upgrade their phones, helping phone manufacturers raise prices and increase profit margins. For example, the price of the Samsung S24 Ultra with a large AI model has increased from $1199 to $1299 from the previous generation, and more subscription-based AI features can be introduced in the future.

Morgan Stanley believes that after global smartphone shipments have declined for three consecutive years, this year is the turning point for global shipments to return to growth. With Apple embedding large models, global shipments are expected to increase by 5% next year Among them, Morgan Stanley expects that the share of AI phones in the smartphone market will increase from 13% this year to 28% next year. It is expected that Apple will account for more than 50% of the global AI smartphone market this year, but as AI phones become more widely popular, Apple's market share in AI phones will decline.

Morgan Stanley predicts that Apple's shipments in 25/26 will reach 250 million units/275 million units, compared to 228 million units in 23/24. The reason for Morgan Stanley's upward adjustment is that AI phones will drive sales in the largest Chinese and North American markets.

At the same time, the EPS for the 25/26 fiscal year is raised to $8.1 and $9.69, while the market's previous EPS expectations were $7.26 and $7.64. Morgan Stanley also raised Apple's target price to $245, corresponding to approximately 29 times PE based on next year's profit forecast.

Foreign institutions believe that Apple's decline in the Chinese market in 2023-2024 is mainly due to Huawei's return to self-made chips. Starting from next year, the share of the global smartphone market will depend on the progress of each company in high-end AI phones. Foreign institutions believe that based on Apple's advantage in applications, Huawei's share will decline next year, Samsung's share will remain stable, and Apple's market share will grow again.

It is worth mentioning that Apple's ability to produce good products is unquestionable, but whether the applications can be smoothly implemented may cause some disputes with regulatory authorities, which may temporarily affect the stock price, similar to the smooth launch of FSD, for example.

Assuming that the AI large models provided by Apple in China are different from the overseas version, and Huawei may also launch its own AI large model phone, the sales performance in China may not be as optimistic. Because it was previously rumored that Apple's domestic version may cooperate with Baidu, or with other domestic large model manufacturers, if it is true, Apple's competitiveness in the Chinese market for large models will be affected.

In addition, Apple stated that due to regulatory issues, it is temporarily unable to launch AI functions in the European market. Apple also mentioned at the conference that the new Siri functions integrated with AI will not appear until next year, and these functions will initially only be available in American English versions, excluding users from other regions.

Apart from regulatory issues, the global smartphone shipments have reached a turning point in the cycle, returning to slow growth is the industry consensus, and which manufacturer provides a better application experience in this round of replacement cycle driven by new technologies is key

2. For the sake of AI, temporarily giving up Vision Pro

Under the logic of the replacement cycle, Apple's stock price has hit a historical high. However, in terms of valuation, the forward P/E ratio is around 30 times. During the previous replacement cycle driven by the sanctions against Huawei, the forward P/E ratio reached as high as around 33 times.

Due to the issues of AI implementation supervision and intense competition mentioned above, the certainty of Apple's future growth is not as good as other tech stocks. This has led to significant market divergence in the price range of 210-220 yuan after Apple's stock price hit a new high. Of course, Apple's buyback amount is significantly ahead of other tech stocks, providing a considerable safety cushion at least.

There are also opinions in the market that do not recognize AI driving Apple's replacement cycle. Take macOS desktop version ChatGPT as an example, which is also a collaboration between Apple and OpenAI, but in reality, it has had no impact on Mac shipments.

The key lies in whether the AI services launched after the collaboration can have a unique competitive advantage.

For example, for large model manufacturers collaborating with Apple, they must have an exclusive partnership with Apple. Otherwise, it is just a way of collaborating by integrating large models, which can be imitated by other phone manufacturers. However, it seems difficult to demand that OpenAI only collaborate with Apple. A more likely way of collaboration is for Apple to design the application, while the actual reasoning relies on the partner's large models and computing power.

It is worth noting that while Apple launched its AI plan, it also adjusted the product strategy of the Vision Pro headset.

Currently, Vision Pro has been on the market for nearly half a year. Although the initial evaluation experience was quite impressive, for more users, Vision Pro is still a very expensive product. At the beginning of its release, it cost about 50,000 to 60,000 RMB in China, and the domestic version is about to be released soon with a price of around 25,000 to 30,000 RMB.

Even though the spatial experience provided by Vision Pro is good, it faces the same problem as the current AI development, which is the lack of a killer application on the application side. Coupled with the initially high price, even if the price is significantly reduced later on, it is still difficult to attract more consumers to purchase.

Due to poor user usage, this year's Vision Pro shipments have decreased from the original 700,000 to 800,000 units to 400,000 to 450,000 units.

Furthermore, Apple plans to launch a lower-priced headset device, expected to be named Apple Vision, with a future selling price reduced from $2,500 to $1,500-2,000 There is a possibility that EyeSight display will be canceled, the function of displaying the user's eyes from the outside will be removed, and a lower configuration chip may be used to reduce the visual effect of combining the virtual and real worlds in the head-mounted display.

Apple plans to popularize a low-cost version to users first, and then continue to develop the high-end Vision Pro. The next generation of Vision Pro is expected to be available by 2026. However, if the low-cost version compromises user experience, even if the lowest price reaches $1500, the price will still be about three times higher than Meta's head-mounted devices.

III. Conclusion

Therefore, even if the hardware experience of the head-mounted device is good, if the application end cannot bring a matching ecosystem, it will all be in vain under high pricing.

In an environment where tech giants are all investing heavily in AI, Apple's investment in AI in the future will certainly not be small. Apple has already abandoned its long-standing car manufacturing plan this year, and the recently launched Vision Pro has disappointed the market. Can Apple continue to invest heavily in AI while nurturing AR/VR devices that are difficult to achieve short-term returns?

Based on these concerns, the market currently believes that a PE valuation of 30 times is reasonable. As for whether it will rise to the highest PE of nearly 33 times in recent years, it will depend on how quickly large-scale models are integrated into smartphones, as other smartphone manufacturers are also doing the same thing