OpenAI is building a phone. Not a concept, not a maybe — according to supply chain analyst Ming-Chi Kuo, mass production is now targeted for the first half of 2027. That’s significantly earlier than the previously expected 2028 timeline.
What We Know About the Device
The OpenAI phone will run a customized MediaTek Dimensity 9600, built on TSMC’s N2P process. The hardware is clearly designed for AI agents: dual NPUs for parallel vision and language processing, LPDDR6 RAM, and UFS 5.0 storage.
The standout feature: an enhanced image signal processor with an improved HDR pipeline — not for prettier selfies, but so AI agents can better perceive the real world visually. The phone is meant to see what you see and act on it.
Jony Ive and the Vision Behind It
The design is reportedly led by Jony Ive — the man behind the iPhone, iMac, and Apple Watch. OpenAI acquired his design studio LoveFrom last year. The idea: move away from individual apps toward a context-aware AI interface that handles tasks seamlessly.
Kuo expects AI agents to fundamentally change how people interact with phones. Instead of opening apps, you simply tell the agent what you need.
The Numbers
If development stays on track, Kuo projects around 30 million units shipped across 2027 and 2028 combined. That’s ambitious for a first product from a company with zero hardware experience.
One potential driver: OpenAI could use the phone to bolster its IPO narrative for late 2026 — hardware makes growth projections more tangible.
My Take
Building hardware is the clearest signal yet that OpenAI wants to grow beyond being a chatbot company. Full control over both the OS and the hardware is the only way to deliver a comprehensive AI agent service — Kuo gets that right. But 30 million units? That would be ambitious for any newcomer, even one valued at $852 billion.
OpenAI hasn’t officially confirmed these plans. But when Ming-Chi Kuo talks about supply chains, it’s worth paying attention.
Sources: MacRumors · 9to5Mac · The Rundown AI