Mark Zuckerberg plans to invest up to $65 billion to advance Meta’s AI initiatives, which includes a data center so massive it could cover a large portion of Manhattan.
The Meta CEO announced that this significant investment will be part of a “massive effort” aimed at enhancing the company’s AI capabilities by 2025. He mentioned a data center in Louisiana that is described as being “so large it would cover a significant part of Manhattan,” in a post on Threads today.
This announcement seems to be a direct response to recent news from competitors regarding their AI data center projects.
Earlier this week, President Donald Trump joined forces with OpenAI’s Sam Altman, Softbank’s Masayoshi Son, and Oracle’s Larry Ellison to unveil Project Stargate, a $500 billion collaboration aimed at constructing extensive AI data centers in Texas and other locations across the nation.
According to city documents reviewed by Bloomberg, the Texas data center could be as expansive as New York’s Central Park.
Meta began constructing its $10 billion AI data center in Richland Parish, Louisiana, last December, with plans for completion extending through 2030.
This facility is one of several that Meta initiated last year to support its open-source large language model, Llama. Alongside Meta, other tech giants like OpenAI, Google, Microsoft, and Nvidia are also investing billions to enhance their AI capabilities.
Zuckerberg anticipates that by the end of the year, Meta will have over 1.3 million GPUs and will significantly expand its AI team.
He stated, “This will be a defining year for AI.”
He expects that by 2025, Meta AI will be the leading assistant for over 1 billion users, Llama 4 will be the top model in the field, and they will develop an AI engineer that will contribute increasingly to their research and development efforts.
AI vs Coding Jobs
Zuckerberg has previously hinted at the potential for AI to take over certain coding tasks within companies. In a recent appearance on The Joe Rogan Experience, he expressed his belief that AI will soon be capable of performing the duties of a “mid-level engineer” at Meta.
He mentioned, “By 2025, we at Meta, along with other companies working on this technology, will likely have an AI that can function as a mid-level engineer, capable of writing code.”
Zuckerberg noted that initially, the cost of running such AI would be high, but efficiency would improve over time. He predicted that eventually, much of the code in their applications, including the AI they develop, would be produced by AI engineers rather than human engineers.
Similarly, Google has been exploring the use of AI for coding tasks. CEO Sundar Pichai revealed last year that over 25% of new code at Google is generated by AI, with human engineers supervising the process.
This shift has raised concerns among industry professionals, particularly recent computer science graduates seeking entry-level positions.
