無程式碼也能輕鬆打造專業LINE官方帳號！一鍵導入模板，讓AI助你行銷加分！
Why is Qualcomm Secretly Collaborating with Multiple AI Giants?
In an exclusive interview with Fortune magazine, Amon explicitly stated that Qualcomm is working with “pretty much all” major AI companies to develop secret devices. He declined to reveal the full list but confirmed it includes OpenAI and Meta. This is not a single-client project but Qualcomm’s comprehensive strategy to dominate the AI hardware supply chain.
The key lies in Qualcomm’s unique positioning: it is one of the few manufacturers capable of providing low-power, high-performance edge AI computing chips. In the smartphone era, Snapdragon chips solidified Qualcomm’s position in the mobile market, but in the AI era, devices will shift from “handheld” to “wearable,” with stricter power and size constraints, making Qualcomm’s technological advantages even more prominent.
TF International Securities analyst Ming-Chi Kuo reported in late April that Qualcomm and MediaTek are jointly designing custom chips for OpenAI for an AI phone, manufactured by Luxshare Precision, with mass production expected in 2028 and annual shipments of 300-400 million units. However, this prediction contradicts Amon’s “non-phone device” statement. Kuo’s updated report on May 5 corrected the direction: the phone chip may be exclusively supplied by MediaTek, with mass production moved up to early 2027.
This change reveals key information: Qualcomm’s role in the OpenAI phone project may be replaced, but it simultaneously dominates more secret devices. Amon’s strategy is not to bet on a single client or device but to become the underlying chip supplier for all AI hardware.
How Will “Your Ecosystem” Replace Smartphones?
The core concept proposed by Amon is “your ecosystem,” a system composed of multiple wearable devices: cameras on glasses capture what you see, earphones precisely receive what you hear, and an AI agent connects all devices in the background, automatically executing tasks.
This is completely different from the current smartphone-centric architecture. Today, all apps, notifications, and operations are concentrated on the phone, but future AI-native devices will be distributed across the body, commanded by an autonomous AI agent. In the interview, Amon gave examples: when you look at a restaurant bill, the AI agent automatically pays; when you see a product, the agent automatically compares prices; when a meeting appears on your calendar, the agent automatically calls to reschedule.
The technical foundation of this vision is edge AI computing. All sensory data—visual, auditory, voice—must be processed in real-time on the device, not uploaded to the cloud, otherwise latency and privacy issues would ruin the experience. Qualcomm’s Snapdragon X series AI platform launched in 2023, and the Snapdragon 8 Gen 3 in 2024, both enhanced on-device AI processing capabilities, a key differentiator from other chipmakers.
graph TD
A[User] --> B[Smart Glasses<br>Visual Perception]
A --> C[Smart Earphones<br>Auditory Perception]
A --> D[Wearable Accessories<br>Context Sensing]
B --> E[AI Agent Engine<br>Qualcomm Chip]
C --> E
D --> E
E --> F[Automated Tasks<br>Payment Booking Search]
E --> G[Cloud Collaboration<br>Training and Updates]Can Smart Glasses Really Replace Phones?
Amon is very optimistic about glasses as a new form factor—“glasses are closest to your eyes, ears, and mouth,” providing the most complete sensory information for AI agents. But he also emphasized that the market won’t have a single winner: “Not everyone wears the same clothes, not everyone wears the same glasses.”
This view aligns with Meta’s Ray-Ban smart glasses strategy and differs from Apple Vision Pro’s positioning. Vision Pro is a closed spatial computing device with high cost and limited use cases, while Amon describes glasses as lightweight, all-day wearable products that integrate with existing lifestyles.
ByteDance’s Doubao assistant has already provided initial validation. In December 2024, the TikTok parent company launched the Doubao assistant on the ZTE-manufactured Nubia M153 phone, where the AI agent can directly operate phone software and navigate apps, perfectly matching Amon’s autonomous agent concept.
But for glasses to replace phones, several key obstacles must be overcome:
| Obstacle | Current Status | Qualcomm Solution |
|---|---|---|
| Battery Life | Smart glasses average 2-4 hours | Low-power AI chip extends to 8-12 hours |
| Heat Dissipation | Continuous computing causes overheating | Advanced process and heterogeneous packaging |
| Privacy Concerns | Camera continuously recording | On-device processing, no cloud upload |
| Price Barrier | High-end smart glasses over $500 | Reference phone chip price curve, below $200 by 2028 |
What Changes Will the AI Device Market See in 2027?
Kuo’s report provides a specific timeline and numbers: the OpenAI phone (or AI device) is expected to enter mass production in early 2027, with annual shipments of 300-400 million units. This scale rivals the peak shipments of the iPhone, meaning the AI hardware market will grow from zero to hundreds of billions of dollars in two years.
But more noteworthy is that the “secret devices” Amon describes are not just one. Qualcomm is collaborating with multiple AI companies simultaneously, indicating the market will see various form factors of AI-native devices:
timeline
title AI Hardware Development Timeline
2024 : ByteDance Doubao assistant launched<br>AI agent concept validation
2025 : Meta smart glasses shipments exceed 1 million<br>Qualcomm Snapdragon X series mass production
2026 : OpenAI secret device design finalized<br>Qualcomm and MediaTek chip competition heats up
2027 : AI-native devices mass production<br>Annual shipments reach 300-400 million
2028 : Smartphone shipments begin to decline<br>Wearable AI devices become mainstream
Qualcomm's stock rose 13% after Kuo's report, and the market has begun to reflect this expectation. But the real impact will occur in 2027-2028 when these devices ship in large volumes, forcing changes in the existing smartphone supply chain, app ecosystem, and even telecom operator models.How Will the AI Agent Business Model Disrupt the Current Ecosystem?
The AI agent described by Amon is not just a technological upgrade but a fundamental shift in business models. Currently, platform companies (Apple, Google) profit through App Store commissions and advertising revenue; in the future, AI agents will directly execute tasks for users, bypassing the app intermediary.
For example, when an AI agent sees you looking at a restaurant bill and automatically pays, it doesn’t need to open any banking app or go through Apple Pay or Google Pay intermediaries. This means existing business models like payment fees, in-app purchase commissions, and ad impressions will be challenged.
ByteDance’s Doubao assistant has already demonstrated this trend: the AI agent directly operates phone software, and users no longer need to manually open apps. If this model becomes widespread, the commercial value of app stores will shrink significantly, and companies that control the AI agent entry point—possibly OpenAI, Meta, or ByteDance—will become the new platform rulers.
| Current Model | AI Agent Model | Affected Parties |
|---|---|---|
| App Store 30% commission | AI agent directly executes tasks | Apple, Google |
| Search ad revenue | AI agent directly gives answers | Google, Meta |
| Payment fees | AI agent automatically pays | Banks, payment platforms |
| Phone hardware profit | Wearable hardware profit | Apple, Samsung |
Can Qualcomm’s Chip Strategy Win in the AI Era?
Amon’s strategy is clear: Qualcomm aims to become the “Intel” of AI consumer hardware. But this goal faces two main challenges: competition from MediaTek and the trend of AI companies developing their own chips.
MediaTek has secured exclusive supply for the OpenAI phone project, showing that MediaTek’s strength in AI chips is not to be underestimated. Kuo’s report indicates that MediaTek’s Dimensity series chips are close to Qualcomm in AI computing performance and are more price-competitive.
On the other hand, OpenAI, Google, and Amazon are all developing their own AI chips. Although OpenAI’s custom chip project is currently outsourced, it may shift to in-house development in the long term. If AI giants follow Apple’s A-series chip route, Qualcomm’s supplier role could be marginalized.
However, Qualcomm’s advantage lies in ecosystem depth. The Snapdragon platform is not just a chip but includes a complete solution with AI engine, modem, and image signal processor. For startups wanting to quickly launch AI devices, Qualcomm’s turnkey solution is more practical than developing their own chips.
Impact and Opportunities for the Taiwan Supply Chain
This trend directly impacts Taiwan’s tech industry. Luxshare Precision has already been named as the manufacturer for OpenAI devices, and TSMC may take Qualcomm’s AI chip orders. But the broader impact includes:
Chip Packaging and Testing: AI wearable devices have stricter size and power requirements, increasing demand for advanced packaging technologies (e.g., TSMC InFO, CoWoS).
Sensor Modules: Smart glasses require multiple sensors (cameras, microphones, accelerometers, gyroscopes), benefiting Taiwan’s optical and sensor supply chains.
Assembly and Mechanical Parts: Wearable device assembly is more challenging than phones, testing and offering opportunities for assemblers like Hon Hai and Pegatron.
Batteries and Charging: Demand for ultra-thin batteries and wireless charging will surge, and the technological layout of Taiwan’s battery module makers will determine competitiveness.
FAQ
What is the secret device Qualcomm is developing with OpenAI?
Qualcomm is working with OpenAI, Meta and other AI giants to develop a secret device that could replace smartphones, likely smart glasses or wearable AI hardware, with mass production expected in 2027.
Will these new devices completely replace smartphones?
The Qualcomm CEO believes the smartphone-centric era will end, with new devices like glasses and earphones connected by an AI agent, but phones won’t disappear immediately; they will gradually become marginalized.
What role does Qualcomm play in this AI hardware wave?
Qualcomm aims to be the core chip supplier for AI consumer hardware, providing low-power, high-performance computing platforms that enable real-time AI processing in glasses and accessories.
What is the market size forecast for the OpenAI phone?
Analyst Ming-Chi Kuo predicts the OpenAI phone will enter mass production in 2028 with annual shipments of 300-400 million units, rivaling iPhone scale, but a recent report moved the timeline to 2027.
Will smart glasses become the next mainstream device?
The Qualcomm CEO is very optimistic about glasses as a new device, as they are close to the eyes and ears, providing the most complete sensory information for AI agents, but the market won’t have a single winner.
