Apple unveiled its reworked Apple Intelligence at this year’s WWDC, fulfilling its long-delayed AI features for iPhones first promised two years ago. Key takeaways are as follows:

1. Apple has partnered with Google to build its new AI system on top of Google’s Gemini models, with 2 on-device variants and 3 cloud-based variants of different sizes. It maintains a strict privacy strategy: a dedicated private cloud infrastructure keeps user data within Apple’s controlled environment, with end-to-end encryption. Apple also opens up for third-party AI integrations, though users will need their own paid subscriptions to use third-party models such as ChatGPT.

2. The updated Siri gains new capabilities including on-screen context awareness, personal context understanding, and cross-application task execution. Core system apps such as Photos and Safari also get new AI-powered features; for example, Photos can now generate new shooting angles for existing images.

3. Apple Intelligence currently places Apple in the top tier of the global AI industry, with an intelligence classification of Level 2 (tool-level). The lack of disruptive innovation in this release led to a post-keynote drop in Apple’s share price, erasing over $230 billion in market capitalization.

This article summarizes key takeaways for smartphone brands from Apple’s latest AI progress:

1. For product R&D direction: The AI smartphone industry as a whole is currently transitioning from Level 2 (tool-level) to Level 3 (assistant-level) intelligence. Consumer expectations for breakthrough AI innovation remain very high. Apple’s post-WWDC $230+ billion market cap decline driven by lack of disruptive innovation underscores that brands must prioritize users’ core demand for AI innovation.

2. For technology roadmapping: Brands can pursue open collaborations when internal R&D hits bottlenecks. Apple addressed gaps in its in-house AI capability by customizing and fine-tuning a model based on Gemini, while retaining its own competitive advantages in tight hardware-software integration and privacy protection that align with user demand for data security.

3. For industry trend assessment: AI has drastically elevated the strategic importance of smart devices. Tight hardware-software integration is the industry’s endgame, and brands with chip and ecosystem advantages will build up increasingly strong competitive moats.

This article outlines key industry opportunities and risks for sellers of AI-related products:

1. Overall industry movement: AI is reshaping smartphones, a shift that has become industry consensus. Beyond traditional smartphone brands, ByteDance has consolidated its former Smartisan (Hammer) smartphone and PICO hardware teams to accelerate its Doubao AI smartphone project, marking the start of full-fledged competition in the AI smartphone track. Sellers can start preparing for AI smartphone-related sales and service business ahead of the curve.

2. Risk warning: AI feature deployment currently remains technically challenging. Apple previously faced a class-action lawsuit over delayed AI feature launches, and ultimately settled the case for $250 million. Sellers should avoid overhyping unpolished AI functions to prevent consumer disputes.

3. Opportunity outlook: The AI smartphone industry is still in the early stage of transformation. Consumer demand for products with innovative AI features is very strong, and AI products with ecosystem integration capabilities will become the main driver of future growth. Sellers can secure early positions in relevant product categories.

This article summarizes key AI transformation insights for hardware manufacturers:

1. Product design and manufacturing requirements: Running large models on-device in the AI era imposes new requirements for smartphone chips, memory, and storage. Apple’s latest M5 Pro chip has already been architecturally redesigned to enhance GPU performance for AI computing and increase memory bandwidth to fit AI workloads. Upstream manufacturers need to keep up with these new hardware requirements, and adjust their product design and production planning to support AI functionality.

2. Business opportunities: The AI smartphone industry is still in the early transformation phase. Top players including Apple, Google and ByteDance are all ramping up investment in AI-powered end devices, which will drive growth in upstream hardware orders. Factories can proactively connect with leading brands for their AI hardware needs, and secure new order share.

3. Transformation takeaways: Tight hardware-software integration is the industry’s endgame in the AI era. Factories can explore deep integrated partnerships with technology brands, participate in end-to-end collaboration from design through production, and boost their own competitiveness amid the AI transformation wave.

This article summarizes key industry insights for AI-focused service providers:

1. Industry development trends: AI smartphones have become the core growth direction of the current smart device industry, and deploying large models on end devices is an industry consensus. Market demand for on-device AI will continue growing, and tight hardware-software integration is the ultimate direction for the industry.

2. Core client pain points: Deploying self-developed large models on end devices remains extremely technically challenging. Apple repeatedly delayed its AI functionality rollout due to technical issues, and ended up paying a legal settlement over the delay. Most device brands face gaps in large model development capabilities, creating strong demand for external technical support.

3. Reference solution directions: Dynamic sparse large model deployment for end devices is already a proven approach. Apple used pruning technology to solve the problem of excessive memory usage by large models. The "base model collaboration + custom fine-tuning + privacy encryption" model also addresses the gap of in-house development capability for many brands. Service providers can refer to these directions to develop market-aligned services.

This article summarizes key insights for AI-focused platform operators:

1. Market demand direction: Most AI end device brands currently lack in-house large model development capabilities. Even leading player Apple had to partner with Google to develop a custom large model. This leaves massive unmet demand for AI technology matchmaking among brands. Platforms can build AI technology supply-demand matching services to meet brands’ collaboration needs.

2. Traffic entry opportunity: Apple has opened up for third-party AI integration, allowing third-party models including ChatGPT and Claude to integrate into its mobile operating system, with users paying third-party providers directly for access. This creates a new traffic entry for AI service platforms to access end users. Platforms can explore collaboration and integration opportunities with mobile brands.

3. Risk mitigation direction: Launching unpolished AI functionality prematurely can trigger user complaints and even legal disputes, as seen in Apple’s previous class-action lawsuit. Platforms need to proactively verify the reliability of AI functions before onboarding AI-related brands and services, to mitigate compliance risks.

This article summarizes the latest industry insights for industry researchers focused on the AI smartphone space:

1. New industry developments: The AI smartphone industry is currently in the critical pre-transformation phase similar to the shift from feature phones to smartphones. Leading player Apple has completed the restructuring of Apple Intelligence, delivered system-level AI functionality built on Gemini, and the overall industry intelligence level is at the L2 tool stage. Internet firms including ByteDance are accelerating their entry into the market, the industry’s competitive landscape is being reshaped, and tight hardware-software integration is widely recognized as the industry’s endgame.

2. New emerging industry challenges: Even leading brands still face technical bottlenecks when developing in-house on-device large models. Apple’s repeated delays to its AI rollout highlight the difficulty of deploying large models to end devices. Meanwhile, consumers and capital markets hold very high expectations for disruptive AI innovation, and the lack of breakthrough innovation will trigger negative market reactions.

3. New business model reference: Apple adopted a "collaborative custom base model + open third-party integration" model. This approach not only addressed gaps in Apple’s in-house AI capability, but also preserved choice for end users. This new partnership model deserves further in-depth research.