OpenAI's highly anticipated "machine-making" plan has made breakthrough progress. Renowned tech analyst Mingchi Kuo pointed out in his latest industry chain research report that the development of OpenAI's first AI agent smartphone has significantly accelerated, with the mass production schedule moving from the previously expected 2028 to as early as the first half of 2027.

This move is interpreted by the market as adding leverage for OpenAI's upcoming IPO process, while also reflecting the increasingly fierce competition in the AI hardware sector. To support powerful local AI computing, the device has chosen a highly competitive underlying architecture: MediaTek is expected to beat Qualcomm and become the exclusive supplier of the phone's processor. It is reported that this customized "Dimensity 9600" chip will use TSMC's most advanced N2P process and is expected to start trial production in the second half of this year.

image.png

In terms of core specifications, this smartphone will be equipped with a dual NPU architecture to achieve efficient heterogeneous AI computing. In response to data privacy challenges in the AI era, the device also introduces "hard-isolated safe box" pKVM and inline hashing technology, ensuring the security of intelligent agent interactions at the hardware level. Additionally, the device's image signal processor (ISP) is listed as a key highlight, aiming to greatly enhance visual perception capabilities so that AI can better understand the real world.

In terms of hardware specifications, the OpenAI smartphone will directly adopt LPDDR6 memory and UFS5.0 storage combination to eliminate storage bottlenecks when running large models. Kuo predicts that if the development process continues as it is, the product's cumulative shipment volume between 2027 and 2028 is expected to reach 30 million units, which is an ambitious goal for a newcomer in the hardware market.