Leading AI company OpenAI is accelerating its expansion into the physical world. In the latest episode of its podcast, OpenAI CEO Sam Altman made a public and clear statement that the company will resolutely develop humanoid robots in-house in the future, while also developing various forms of robots and specialized robots for data centers.

Altman pointed out that since the physical space of human society is perfectly adapted to the human body structure, the humanoid design naturally became an important direction in robot development. Earlier this June 1st, OpenAI had already posted job postings related to OpenAI Robotics on social media platforms, recruiting full-stack hardware, systems, and machine learning engineers. He revealed that the company's early world simulation research project has developed rapidly over the past year and has now evolved into OpenAI Robotics, led by Aditya Ramesh. At present, its core underlying logic lies in the deep collaborative design of robot hardware research and machine learning. In terms of application rhythm, in the short term, OpenAI focuses on developing robots that can assist technical workers in building future infrastructure; in the long run, the company envisions that everyone will have a personal robot capable of meeting various lifestyle needs.

In addition to the grand plans in the field of embodied intelligence robots, OpenAI's first consumer-level AI hardware product is also about to be launched. The device was designed by former Apple Chief Designer Jony Ive, with an internal code name "Gumdrop" (Candy). It is a screenless portable AI terminal that emphasizes a "more peaceful" user experience than smartphones, with a shape similar to a donut, about the size of a hockey puck. The design concept allows users to hold the device with one hand and move it around the house. It features functions such as handwritten note transcription and real-time voice interaction, and can directly connect to ChatGPT to provide services like summarization, translation, and mind map generation. Its price may exceed $300 and is planned to be officially released in 2027.

To advance its ambitious hardware matrix, OpenAI has not only successfully recruited hundreds of top hardware engineers from tech giants such as Apple and Google, but also entered into a deep collaboration with Foxconn to promote prototype production, with the first mass production plan already scheduled. According to reports, OpenAI's hardware team is currently developing approximately five different products in parallel and hopes to start with this speaker-shaped screenless terminal, turning it into a physical embodiment of ChatGPT.

However, behind the rapid progress in hardware, OpenAI also faces considerable legal challenges. Apple formally filed a lawsuit against OpenAI this July, accusing it of stealing trade secrets and obtaining information about unreleased Apple products through coordinated actions. Subsequently, in August, Apple again took action, requesting a federal judge to issue an order requiring OpenAI to immediately stop using and return all relevant confidential materials. In response, OpenAI publicly stated that Apple's allegations were unnecessary and based on incorrect information, emphasizing that the company focuses more on breakthroughs in cutting-edge innovation technology.

In the capital market, OpenAI is also preparing major moves. The company's CFO, Sara Ferial, recently revealed that the company plans to complete its initial public offering (IPO) in 2027. If business growth accelerates significantly, the listing timeline could be further advanced.

From the financial fundamentals perspective, OpenAI's revenue in the second quarter reached $6.7 billion, an 18% increase compared to the first quarter. Looking back at previous performance, its annualized revenue for 2025 exceeded $20 billion, representing a growth of more than 230% compared to $6 billion in 2024. However, industry comparison data shows that facing strong competitors such as Anthropic (whose latest quarterly revenue has surpassed $11.5 billion, with an annualized operating rate exceeding $46 billion and achieving quarterly profitability), OpenAI still faces the reality of slowing revenue growth and expanding operational losses.