iPhone 16 Sales Struggle: Is AI Not the Savior of Smartphones?

iPhone 16 Sales Struggle: Is AI Not the Savior of Smartphones?Recently, Barclays analysts Tim Long and George Wang noted in a report that Apple may be cutting production for the fourth quarter due to lower-than-expected demand for the iPhone 16 series. Supply chain data shows that the iPhone 16 series has seen order cuts of around 3 million units

iPhone 16 Sales Struggle: Is AI Not the Savior of Smartphones?

Recently, Barclays analysts Tim Long and George Wang noted in a report that Apple may be cutting production for the fourth quarter due to lower-than-expected demand for the iPhone 16 series. Supply chain data shows that the iPhone 16 series has seen order cuts of around 3 million units. Barclays believes that the shortened delivery time for the iPhone 16 series also indicates weak demand for the new phone, primarily due to a lack of compelling new features and an underdeveloped AI app ecosystem.

Supply chain insiders point out that smartphones need to be redesigned to accommodate AI applications, a process expected to be completed in 2026 or 2027. In recent years, the mobile and PC industries have been transitioning to AI, with AI phones and AI PC products emerging. However, the AI capabilities of almost all these products feel like gimmicks, making it hard to differentiate them from regular phones and PCs. Why have AI features, which phone manufacturers held such high hopes for, failed to become core competitive advantages for new phones and attract existing users to upgrade?

AI Phones: Features vs. Experience

Citigroup argues that the underwhelming sales of the iPhone 16 series stem from the absence of Apple Intelligence, its key feature, which is yet to launch. They predict that the launch of Apple Intelligence on October 28th, the major Siri update next year, and the launch of the iPhone 17 series will attract a significant number of iPhone users to upgrade.

However, these analytical agencies and smartphone and PC manufacturers promoting AI features as major breakthroughs overlook a fundamental issue: current AI experiences have little to do with a device's AI performance or features. In recent years, many phone manufacturers have touted their built-in AI applications for functions like document extraction, image generation, writing, and coding. However, these functions are all cloud-based, not executed on the phone itself. A flagship phone costing several hundred dollars can achieve these same functions, and so can mid-range models costing less than half as much. Even if a phone doesn't offer related AI applications, users can leverage third-party AI apps like ChatGPT, Bard (overseas), and Doubao, Wenxin Yiyin, and Xunfei Xinghuo (domestic) to achieve these functionalities through apps or web browsers.

Whether it's generating images, text, or assisting users in writing code, since all of these can be achieved in the cloud with numerous apps readily available, it's difficult for phones to differentiate themselves from competitors using AI features. This makes it hard for them to become core competitive advantages and easily leaves consumers with the impression that they are just gimmicks.

Apple Intelligence, set to launch on October 28th, boasts eight features: AI comprehension, AI summarization, article summarizing/rewriting/generation, AI replies, AI recognition, AI filtering, AI transcription, and AI assistant. The majority of these functions can be achieved using third-party apps. Relying solely on these commonly supported AI features, the iPhone 16 series struggles to offer something unique. Its impact on sales is likely to be limited, requiring Apple to continuously add more features to Apple Intelligence.

Smartphone AI is Not a Gimmick, Just an Ecosystem and Feature Set Needing Refinement

For this reason, smartphone manufacturers are searching for a breakthrough, exploring directions like on-device AI and AI Agents. OPPO, Honor, and other manufacturers have launched flagship models equipped with 70-billion-parameter on-device large language models. At the recent IFA 2024 in Berlin, Honor showcased the capabilities of AI Agents and announced that the Magic7 series would be the first to feature this technology.

Compared to cloud-based AI, on-device AI has two main advantages: First, it offers shorter interaction latency, unaffected by network fluctuations. Second, data is computed on-device, effectively protecting user privacy. As they introduce on-device AI, major phone manufacturers emphasize data security and the ability of on-device AI to generate large language models based on user habits, creating AI assistants that better align with user preferences and provide more convenient assistance for various tasks.

Of course, on-device AI also has notable disadvantages: With limited computing power on the phone, it's suited for handling real-time, low-complexity tasks with low performance demands. However, it may struggle with high-performance training tasks and some inference tasks. The massive parameter scale and energy consumption issues are also key obstacles to the development of on-device AI.

To address the issues of parameter scale and energy consumption, Apple introduced Apple On-Device large language models with 3 billion parameters; Google DeepMind released Gemma22B, with 2.6 billion parameters, further reducing resource requirements and improving operational efficiency.

Apple's A-series chips and Qualcomm's Snapdragon chips are both enhancing AI computing power, laying the foundation for the advancement of on-device AI. Chip AI performance and on-device AI development are mutually supportive: the former provides the performance foundation, creating space for the latter to flourish; the latters improved features unlock the full potential of the former.

AI Agents represent the target form for on-device AI. Users simply need to give instructions, and the AI Agent will independently think and execute the necessary actions, eliminating the need for users to issue commands for every step. At IFA 2024, Honor showcased an AI Agent feature that automatically disables auto-renewal with a single command. App auto-renewal buttons are usually buried in four or five layers of menus, making them difficult for users to find. Honor's AI Agent, however, can instantly help users disable them with a simple voice command, highlighting the power of AI Agents.

Apple emphasized at its June developer conference that the new version of Siri will have screen awareness, enabling it to understand the content a user is viewing and provide corresponding suggestions. It will also integrate deeply with third-party apps for easier user interactions. The focus of Honor and Apple's AI features indicates that deeper understanding of user instructions and cross-app operations will be essential capabilities for future phone AI.

Understanding user instructions is crucial for AI to accomplish complex tasks, while enabling cross-app operations is key to enhancing user experience. In the simpler automotive ecosystem, some carmakers have achieved "visibility equals verbal control," meaning users can select or perform actions using voice commands on any text displayed on the screen. Honor goes even further, enabling direct voice command control of options within four or five-level menus. However, considering the vast number and complex functionalities of phone apps, it's likely that only a handful of apps deeply integrated with phone manufacturers will initially be able to achieve full control through voice commands.

AI Phones Accelerate: Science Fiction Becomes Reality

While AI functionalities may not yet serve as core competitive advantages for smartphones, we've witnessed progress in phone AI, from on-device AI to AI Agents. As supply chain insiders mentioned, smartphone hardware needs improvement to align with the AI era's application ecosystem. Progress in the AI field requires collaboration between the supply chain and smartphone manufacturers, leaving consumers to wait and invest real money to support AI products.

The promised functionalities of Apple Intelligence and Honor AI Agent provide a glimpse into how AI will enhance future phone experiences. More importantly, the cross-app operations and the ability to call upon AI anytime and anywhere are difficult for third-party AI apps to achieve. These features could become core competitive advantages and sales drivers for future flagship phones, and could also help drive growth for the iPhone through Apple Intelligence.

It hasn't been long since Apple, Qualcomm, and other phone manufacturers fully committed to AI, and the iPhone 16 series marks Apple's first phone specifically designed for Apple Intelligence. It's estimated that global phone AI technology will mature around 2026, at which point AI Agents will be able to comprehend human natural language and perform deep cross-app operations.

Since the concept of AI was first introduced in 1956, nearly 70 years have passed, with hardware and software technologies steadily improving. Phones, essential to everyday life, are poised to become the optimal vehicles for AI. From feature phones to smartphones, and now from smartphones to AI phones, the mobile industry is entering a new era. While the high level of concentration in the phone industry makes it difficult for new giants to emerge, some brands could fall behind.

Mediatek and Qualcomm will be releasing new flagship chips in October, followed by the launch of a wave of new Android flagships. AI, undoubtedly, will be the central focus of product differentiation. Challenges and opportunities lie ahead for all phone manufacturers, as competition intensifies. It's time to test their strength.