Meta (META.US) hires 4 more OpenAI core researchers, spends money to attract talents to accelerate catching up in the AI competition.

Meta (META.US) is investing billions of dollars to take the lead in the artificial intelligence competition, building the necessary data centers to develop and drive high-end, large language models. Now, the company is investing billions of dollars in recruiting top talent and advanced technology to gain an advantage in the AI battle. Reportedly, the US tech giant Meta has poached four AI researchers from OpenAI. The four researchers poached by Meta are all recognized as technical leaders within OpenAI, including Jiahui Yu, the lead of the o3, o4-mini, and GPT-4.1 projects, Hongyu Ren, creator of the o3-mini and o1-mini models, Shuchao Bi, leader of the post-training team for multimodal models, and Shengjia Zhao, a key contributor to the GPT-4 and o1 projects. Previously, it was reported that Meta had already hired Lucas Beyer, Alexander Kolesnikov, and Xiaohua Zhai in their latest talent recruitment effort, all of whom had previously worked at OpenAI's Zurich office. OpenAI CEO Sam Altman stated that Meta CEO Mark Zuckerberg had promised to offer up to $1 billion in compensation to attract them to join their AI project. Meanwhile, reports suggest that Meta is negotiating with Safe Superintelligence CEO Daniel Gross and former GitHub CEO Nat Friedman to incorporate them into their planned Superintelligence Lab. Initially, Zuckerberg had planned to acquire the entire Safe Superintelligence company but was rejected by co-founder Ilya Sutskever. These moves come after Meta's investment of $14.3 billion in the AI startup Scale AI and hiring its CEO and co-founder Alexandr Wang. Furthermore, Meta had expressed interest in acquiring Perplexity AI but failed to reach an agreement on the deal. Forrester analyst Mike Proulx stated, "Meta is doing this because they want to win in the AI competition, plain and simple. Right now, AI is everything." All of these actions were triggered by Meta's decision to postpone the launch of their Llama 4 Behemoth AI model. The company reportedly delayed its release until the end of this year's autumn, fearing that the upgrade might not be significant enough compared to previous versions. Gene Munster, managing partner at Deepwater Asset Management, commented, "I think this mainly has two meanings: first, it shows that Meta is facing challenges, and second, it shows that Zuckerberg is not satisfied with it." For Meta, the goal is clear: to attract as much fresh talent as possible to drive the development of their AI projects and take the lead in the AI competition. Meta's AI Initiatives Meta's efforts to dominate the AI field differ from its main competitors like OpenAI, Alphabet Inc. Class C, and xAI. The company does not keep its AI models closed but offers them as open-source software for developers and companies to use on their own. Meta imposes certain restrictions on how users can utilize its models. For example, the company requires businesses to seek permission from Meta if their product exceeds 700 million monthly active users. Ultimately, Meta aims to have as many users as possible utilize and develop their products through their AI models. Why doesn't Meta charge everyone who wants to use their software? Because every time a company modifies their model, Meta gains valuable data for future improvements. Meta is not very interested in selling the rights to use their models. The company primarily uses its AI technology to support its advertising and content recommendation services, unlike Microsoft Corporation (MSFT.US), which sells its AI services as part of its office software package. Meta's CFO Susan Li stated in the company's recent earnings call that since introducing the Llama algorithm into its recommendation system at the end of last year, the user usage time for the Threads application has increased by 4%. Additionally, Meta relies on its AI models to provide intelligent support for its hardware products, including their Ray-Ban Meta smart glasses and other devices based on AI technology in the future. Proulx explained, "Meta is taking these actions because they have a lot of funds. With this funding, even if they can't develop it themselves, they can use it to acquire the talent and capabilities needed... to surpass their competitors." However, Meta is not the only company eyeing the new AI players in Silicon Valley. Apple Inc. (AAPL.US) is also considering acquiring its own AI company as the iPhone maker seeks to improve its position in the AI field. Apple Inc. originally planned to launch an AI-driven version of Siri earlier this year but has delayed the release until the second half of the year due to development difficulties. Reportedly, Apple Inc. also considered acquiring Perplexity AI. Additionally, Apple Inc.'s competitor Samsung is planning to incorporate Perplexity's technology into its products. A spokesperson for Perplexity stated that the company is not currently aware of any ongoing or future acquisition discussions. They added, "The world's best original equipment manufacturers want to provide users with the best search services and the most accurate AI technology, which is the advantage Perplexity possesses." For Meta, success will depend on its ability to attract suitable talent from top AI companies to join their AI project and establish the company as a leader in the AI competition. If they fail to do so, this social media giant may find itself falling further behind without a clear path forward.