Apple’s foldable bet isn’t the screen — it’s the AI

Weeks after Apple unveiled the iPhone Duo, the model’s crease, hinge, thermal performance and durability came under scrutiny.

Those concerns are real. Reports that hinge production is limiting output have prompted GF Securities analyst Jeff Pu to cut his 2026 shipment forecast from 7 million to 6 million units.

But the more important question is not whether Apple can make a good foldable. It is why it entered the race now — and what it is really betting on.

Seven years of waiting

Samsung and Chinese phone makers have spent roughly seven years learning how to build foldables. Samsung’s first Galaxy Fold in 2019 exposed how immature the technology was. Since then, Huawei, Xiaomi, Oppo and Vivo have spent years refining hinges, flexible displays, ultra-thin glass and manufacturing.

By 2026, the core problems appeared largely under control. Apple’s strategy looked obvious: let others absorb the cost of experimentation, then enter when the supply chain is mature. Market researcher TrendForce expects Apple to capture 24.8% of the global foldable market this year, roughly matching Huawei.

It looks like a familiar script: Chinese manufacturers build the supply chain, while an American company arrives later to capture the premium market. But early production problems tell a different story.

These are engineering problems, not procurement problems. Apple cannot simply wish them away, just as Chinese and Korean brands could not when they first entered the foldable market.

Early reviews have raised concerns about thermal performance, screen vulnerability, multitasking and app compatibility — problems the industry has spent years trying to solve. That raises a more uncomfortable question: what exactly did Apple gain by waiting?

Apple’s traditional advantage is arriving late but making technology feel finished. That expectation is particularly strong for a US$1,999 foldable. A company that pioneers a category can ask consumers to tolerate rough edges. A company that arrives seven years later has to show why its exorbitantly priced product is materially better.

The Duo does not yet have an obvious answer.

The real bet is AI

Apple’s real bet is not the foldable display. It is what the display makes possible in the age of agentic artificial intelligence (AI). The conventional smartphone assumes the human operates the apps. An AI-native smartphone reverses that relationship: the user gives an instruction, while the system finds the relevant apps and information, executes a sequence of actions and returns the result.

That is where a foldable becomes strategically interesting. Its larger display can show source information, AI-generated results and the applications being used to execute a task — all at the same time.

The screen is therefore not the destination. It is as much the workspace for AI as a gateway to a new AI-native way of life and work. That makes Apple’s hardware issues more important, not less. Agentic AI requires reliable hardware and software as agents move between applications.

Apple’s own vision for this future rests on Siri AI, which points towards a more proactive assistant with personal-context awareness, onscreen understanding and the ability to take actions across apps. But the system remains in beta, and its tightly controlled app ecosystem also limits how freely agents can act compared with some Chinese approaches.

Chinese phone makers are taking a more aggressive route: embedding agents deeper into their operating systems and the services users rely on every day. In the second half of September, Vivo’s OriginOS 7, Oppo’s ColorOS 17, Honor’s MagicOS 11, Xiaomi’s HyperOS 4 and Huawei’s intelligent-agent framework all shifted attention towards system-level agents capable of executing tasks across applications.

Honor says its Yoyo agent can execute long-horizon tasks involving more than 100 steps. Chinese manufacturers are increasingly rolling out agents that can read information from one app, make a decision and act in another.

The approach has also exposed a fundamental problem. In December 2025, ByteDance put its Doubao AI agent into a limited-run Nubia smartphone, allowing it to operate apps such as WeChat and Meituan on users’ behalf, but major app platforms soon restricted what it could do inside their services.

The lesson is clear: an AI agent can be technically capable of controlling an app without having the authorisation to do so.

Who controls the agent?

Chinese smartphone makers are betting on execution by in-house AI agents — making the operating system understand intent and coordinate actions across services.

Apple is setting its sights on ecosystem control. Rather than owning every AI model, it is positioning the iPhone as a platform for multiple AI providers. Apple already supports ChatGPT through Apple Intelligence and is opening its software platform to additional third-party services such as Claude and Gemini.

Apple does not necessarily need the best model. It needs to control the device, operating system, permissions and user experience through which models act.

Once an agent starts booking travel, sending messages or making purchases, however, the question of authorised access becomes more disconcerting. Users will care less about how many tasks an agent can complete than what it is allowed to do.

The iPhone Duo’s biggest challenge is not that Apple has failed to build the world’s best foldable. It is that Apple is asking a product with unresolved manufacturing and thermal issues to carry a much bigger strategic narrative about the next generation of edge computing. 

But that does not mean Apple’s software advantage has disappeared. Its new Siri AI brings personal-context awareness, onscreen understanding and cross-app actions to a broader range of iPhones, giving Apple a strong foundation for the agentic-AI era.

In China, however, Siri AI remains unavailable, weakening Apple’s software proposition in one of the world’s most important smartphone markets. That leaves Chinese phone makers more room to push system-level AI agents — including on their foldables.

The China market

IDC data show Huawei and Honor held 60% and 21% of China’s foldable market respectively in the first quarter of 2026. The stakes are higher now, but years of accumulated engineering experience do not disappear overnight. And the AI-agent battle is only beginning.

Each contender brings a different strength: Apple has the ecosystem, Samsung has the longest experience with foldables and Chinese brands are pushing agents deeper into the operating system and local apps.

Their weaknesses are different too: Apple faces a slower AI rollout, Samsung depends heavily on Google’s AI stack, while Chinese players face a more fragmented app landscape. 

For now, I would place my bet on China’s phone makers in the agentic AI race. They are already turning the phone from something we operate into something that can operate for us, namely, a workplace assistant.

That is ultimately what the Duo is competing for. The next smartphone battle will not only be about the thinnest hinge or smartest chatbot, but also about turning user intent into actions that actually get done. Chinese makers may have a head start, but Apple’s control over the device experience gives it the leverage to close the gap.

Related: Winning the AI race without an OpenAI | Huawei Mate XT vs iPhone 16: Why Huawei is leading in China