虎嗅

**OpenAI's Hardware Roadmap Revealed: The First Device Has No Screen, and Mobile Phones Will Be Mass-produced in the First Half of 2027**

原文:OpenAI 硬件路线图曝光:第一台硬件没有屏幕,手机2027上半年量产

Summary of Key Points

OpenAI is building a hardware ecosystem centered around “AI companions,” which includes screenless speakers (to be released in early 2027 for $200-$300), AI-powered phones (mass-produced in the first half of 2027), as well as future products such as glasses and desk lamps. The core of these devices is not just functionality, but the ability for AI to continuously monitor the user’s living environment through sensors like cameras, thereby becoming proactive “intelligent partners.” However, they currently face challenges from Apple’s legal actions, supply chain issues, and the critical test of consumer acceptance regarding privacy.

Detailed Analysis

1. These devices are not “toys” – they are AI’s “eyes and ears”

All of OpenAI’s hardware solutions aim to enable AI to “see” and understand users’ lives. Traditional smart speakers only have “ears” (microphones) and know nothing unless you speak to them; in contrast, OpenAI’s devices come equipped with cameras. The speakers have built-in cameras, and the phones feature dedicated chips for processing camera images as a key specification, along with two separate AI processors for handling visual and language tasks.

In simple terms, AI was once “blind” and had to wait for users to initiate interactions; now, with cameras, it can observe on its own. For example, it might remind you to be careful with a knife while you’re chopping in the kitchen or dim the lights when you yawn on the sofa. However, this means users must accept that a device with a camera is watching their home, and privacy is a significant consideration.

2. Screenless speakers: not just speakers, but “living” AI companions

This speaker, designed in collaboration with a former Apple designer, defies convention: it has no screen, is rechargeable and portable, and includes moving mechanical components (such as the ability to turn towards the user). It is positioned as an AI companion that lives in your home, driven by ChatGPT, develops a personality over time (knowing your favorite music and sleep patterns), and initiates conversations with you. Unlike traditional speakers that serve only as alarms or music players, its breakthrough lies in its proactivity—it can provide services without user intervention. However, the price of $200-$300 is much higher than regular smart speakers, and users must trust a device with a camera in their home.

3. AI phones: not just phones, but dedicated carriers for AI agents

OpenAI’s phones are not intended to compete directly with Apple or Android; instead, they aim to provide a closer connection between users and AI agents. Their design focuses on enabling AI to better understand the world (e.g., using enhanced HDR cameras in low-light conditions and dual AI processors). The manufacturing is done by Foxconn (Apple’s partner), the cameras are provided by Sunny Optical (also a supplier for Apple), and the chips use MediaTek’s latest technology (on the same manufacturing node as Apple’s A20 chip). Although the production target of 30 million units per year (for a mid-range brand) is modest, the goal is to verify the feasibility of an “AI-first” system. The biggest challenge is how users will interact with services like banking, delivery, and social media without dedicated apps, which OpenAI plans to solve through subscription models.

4. Apple is the biggest obstacle

Apple has recently sued OpenAI for stealing trade secrets (such as metal surface treatment techniques) and obtained a restraining order, potentially preventing the release of these devices. Additionally, Apple offers generous retention bonuses to its employees, making it difficult for suppliers to collaborate with OpenAI, which has slowed progress. Interestingly, Apple’s own roadmap closely resembles OpenAI’s: smart glasses, camera-equipped AirPods, AI pendants, and home hubs. Both companies are competing in the field of using AI to understand user behavior, and OpenAI’s team and supply chain have significant Apple influences, which Apple certainly does not want to see replicated.

5. The biggest challenge is convincing consumers to pay

OpenAI can handle chip production and supply chain issues, as well as legal battles, but the ultimate question is whether consumers will buy these products. For example, would you be willing to place a $200-$300 screenless speaker in your living room, fearing it might be eavesdropping on your activities? Traditional smart speakers have been widely adopted for years but often remain unused; OpenAI’s devices are more radical, requiring both financial investment and the sacrifice of privacy. This is the true test of whether consumers will accept AI as a integral part of their daily lives.

In Conclusion

OpenAI’s hardware initiatives are not about selling products so much as transforming AI from a feature in phones into a companion that is always present. To achieve this, they must first overcome Apple’s barriers and gain consumer trust. Whether they will succeed will be revealed in 2027.