Meta’s hiring in Artificial Intelligence has attracted global attention, due to the rumor -rumor offering. Reports show that CEO Mark Zuckerberg has increased salary packages up to Rs 1,600 crore on some new fare. However, in a recent episode of the information TTV, Zuckerberg clarified that large salaries play a role – and said that the reports reported exaggerated – they are not only the top AI talent from Open, Google Deepmind and Anthropic, joining the meta. According to him, the money is only a part of the equation.

“Historically, when I was recruiting people in different parts of the company, people are like, ‘Okay, what is going to happen?” “Zuckerberg said. Here, people say,” I want the lowest number of people and the most GPU is reporting. “

GPU, or graphical processing unit, are necessary to develop powerful AI models. These high-demonstrations, especially in-demand H100s of NVIDIA, are used to train and run large language models and other complex AI systems. According to Zuckerberg, “Most of the most calculations per researcher are certainly a strategic advantage, not only to work but also to attract the best people.”

The current AI recruitment drive of the meta is one of the most aggressive silicone valley. The company is allegedly paying salaries between Rs 800 crore to Rs 1,600 crore ($ 100 million to $ 200 million) for allegedly wooing elite researchers of rivals such as OpenI, Apple, Google Dipmind and Anthropic.

The new recruitments are trapit bansal, a former OpenIAI scientist known for his work on the O-series model, and Apple’s ruins pang. Both are now part of the newly formed Superintending Lab of Meta, headed by East-Githab CEO Nut Freedman and former Scale AI CEO Alexandra Wang. The lab is already shaping to have a powerhouse, with at least 11 top researchers from Google and Openai. These include names like Lucas Bayer, Johua Jhai, Jack Rai, and Johankwik Google, as well as Zee Lynn, Shengzia Zhao, and Jiahui from Openai to Ziahui You. Their goal: To help create a Meta Artificial General Intelligence (AGI), a system that can learn, and understand like a human being.

Meta is supporting this talent with heavy infrastructure investment. Zuckerberg announced earlier this week that the company is building an AI superclaster named Promthius, which is expected to go online by 2026. It is part of a comprehensive AI build-out, including data centers such as hyperians that will be up to five gigawatts, one of the largest AI infrastructure declared so far.

To speed up deployment, Meta has started building temporary data centers inside the tents similar to the Tent’s Makshift “Tent Factories” during the production crunch of tents. These prefab structures will allow AI training to start as soon as possible, even if they are not correct for extreme conditions. The ramp-up comes after the alleged lukewarm response to the Meta’s Lama 4 model earlier this year. In response, the company has resumed its roadmap, has invested $ 14 billion in Scale AI for high quality training data and has exposed its wallet widely for the best mind in AI.