According to news on February 26, AI technology executive Pang Ruoming, who has just joined Meta less than 7 months ago, has officially joined OpenAI. This Chinese AI technology tycoon, who was poached by Meta with a multi-year salary package of more than US$200 million, a Chinese AI technology tycoon with a franchise fee comparable to Ronaldo, has become the latest key talent to be attracted by OpenAI.


Back to July 2025.

At that time, Meta was accelerating the promotion of "Superintelligence Labs" (super intelligence laboratory), promoted by Zuckerberg himself, with the goal of AGI.

In order to complement its model training and system architecture capabilities, Meta has high-profile poachers from companies such as Apple and Google.

Pang Ruoming is one of them.


Multiple media reports stated that Meta provided it with more than$200 million multi-year compensation package, including cash and equity incentives. This number is considered to be the top strategic level of talent allocation within the technology industry.

At Meta, he is responsible for AI infrastructure, ultra-large-scale model training systems, distributed computing and computing power efficiency optimization.

This type of position is not responsible for external product release, but determines whether the model can be "implemented on a large scale."

Being able to take on this important task, Pang Ruoming also had a lot to do with Brand.


He graduated from Shanghai Jiao Tong University with a bachelor's degree and a doctorate from Princeton University.

After graduation, he worked at Google for about 15 years, participating in large-scale system and infrastructure development. Then he was responsible for the basic model team at Apple and managed a hundred-person engineering system.

Yes, he is in charge of Apple Intelligence.


Later, he joined the Meta Super Intelligence Laboratory in 2025 and received the sky-high annual salary package just mentioned.

But why did he leave after 7 months? You know, strategic job-hopping for senior-level technical personnel usually takes an annual cycle, but Pang Ruoming left Meta in just 7 months.

According to Reuters, OpenAI’s recruitment lasted for several months and was a key introduction.

Now OpenAI is promoting larger-scale multi-modal models, and inference models and agent systems require higher computing power efficiency. At the same time, the scale of training continues to expand, and the dependence on the capabilities of the underlying system increases.

And Pang Ruoming's core ability falls at this intersection.

OpenAI is naturally unwilling to give up such an opportunity to poach employees.


This offer also sends a clear signal:

At this stage, computing power scheduling, distributed optimization, cost control, and training stability are becoming new strategic highlands.

Previously, Meta took the initiative to poach people from Apple, OpenAI, Anthropic and other companies, and OpenAI also recruited Meta's core members in reverse.

The salary level of top technical personnel continues to set new records, and the turnover of core system personnel will become more frequent.