EP696 | 🎖️

The battle for the foldable iPhone is about a mature experience, not who launches first

The host believes Apple$AAPL's next-generation Pro Max upgrade largely follows the existing path. The real product worth watching is the foldable iPhone Duo. Its price is only slightly higher than the market originally expected.

If fully specced out, it could still cost more than NT$100,000, but products like this are not sold on conventional value for money. They are aimed at users with a clear need who are willing to pay for a better experience. The host also says he would buy one immediately, and therefore believes sales could outperform the market's original expectations.

He does not think “who makes a foldable phone first” is the important question. Google$GOOGL, Samsung, and other Android manufacturers may already have tried similar designs, but technology products are ultimately judged by who can make a feature mature and widely usable, not by who gets the head start. The host uses Nintendo7974.T's hardware and gaming experience as an analogy: its hardware specifications may not be the most advanced, yet strong content and intuitive usability can create a competitive advantage.

Apple's advantage lies in packaging complex functions into controls that ordinary people can understand, while integrating hardware and software to reduce friction. Even if a foldable screen still shows a faint crease, consumers may accept a higher price as long as the overall touch response, switching between two screens, and everyday workflows feel natural. The host therefore expects the market share of foldable phones to rise in this generation, but China's consumption downgrade remains the key point to monitor and verify.

The hinge supply chain may have share, but not necessarily pricing power

The component design of foldable phones differs from that of ordinary phones, and the hinge is one of the key parts. Taiwan's Shin Zu Shing3376.TW is viewed by the host as one of the main suppliers, while U.S.-based Amphenol$APH also holds an important position in related components. Shin Zu Shing accounts for a meaningful share of the product. Its hinges can support multiple angles, not just fully open or fully closed positions, and can also form a tent-like shape. Behind this are technologies involving structure, durability, and support at different angles.

However, Shin Zu Shing's share price and earnings are not easy to project directly. One reason is that it must share the order value with Hon Hai2317.TW. The host originally thought the first generation of the new product would bring excellent gross margins, but industry sources suggest that Apple's bargaining pressure remains strong. This shows that winning a large order does not mean a supplier will obtain a proportional share of the profits.

The host compares Apple's and Tesla's$TSLA vertical integration. Both companies tightly control components that are critical, scarce, or especially valuable. This may lead suppliers to invest first because of the order volume, only to face the risk of lower prices and customers developing alternative suppliers later.

If a supplier becomes overly dependent on Apple, it could eventually become a business with “revenue but no pricing power. ”

Apple's platform value may matter more than any single hardware cycle

The host does not plan to treat the foldable-phone supply chain as a major consumer-electronics investment theme. Compared with AI, even growth in consumer electronics may not attract the most capital. Apple's suppliers may also face stronger cost pressure in the second and third generations of the product. For example, as Apple continues developing its own chips, the value Qualcomm$QCOM can capture may gradually shrink.

Even so, he remains positive on Apple's platform stickiness. Even if components such as memory become more expensive, Apple is better positioned to preserve its gross margins while attracting new customers at relatively acceptable prices. Bringing users into its hardware, operating system, and services ecosystem first, then considering how AI applications can connect to it, may offer more long-term value than simply chasing the specifications of a particular phone generation.

The host also notes that large language models for everyday uses may gradually become commoditized. Looking up recipes, asking about bicycle parts, or planning a hiking route does not require the most advanced model. Platforms that control the user entry point and distribution channels may capture more value instead. This is why he believes Apple, even if it is not among the earliest movers in the AI era, may still maintain its competitiveness by retaining existing customers and attracting new ones.

AI data-center demand is beginning to face constraints from power and local politics

Oracle$ORCL's recent comments drew the host's attention to the financing structure behind AI infrastructure. Companies can prepay expenses or bring their own hardware, Oracle can arrange financing for customers, and suppliers may provide funding for equipment. This model can expand quickly while demand continues to rise, but it may also magnify underlying demand and create an appearance of “inflated growth.” Once the cycle turns, financing pressure could contract at the same time.

AI hardware depreciation may also not be occurring as quickly as bears expected. Four-year-old cards have even risen in price. But that does not mean the risk has disappeared: when demand shifts from strong to weak, equipment prices could still fall rapidly. The host believes Oracle's 30% year-over-year revenue growth and more than 660 billion in remaining performance obligations, or RPO, are impressive overall. Still, in a volatile market, positive news does not necessarily push the share price higher.

The larger bottlenecks are power, transmission lines, and construction permits. Some U.S. pipeline projects have been rejected, while local governments may require companies to provide substantial credit guarantees. This shows that AI data centers have moved from being a corporate investment issue into the realm of local politics. Politicians such as Bernie Sanders and some voters oppose data centers because of concerns about electricity, water resources, the environment, and how benefits are distributed. Even if companies believe the evidence supports construction, voters may still oppose it if they feel only a small group will profit.