Joining a flurry of billion dollar announcements by global tech giants for AI investments, Meta will invest around $65 billion to boost its artificial intelligence (AI) infrastructure this year. Meta Platforms CEO Mark Zuckerberg announced this on Friday (January 24). The parent firm of Facebook, Instagram, Threads, and WhatsApp will allocate this investment mainly for the development of a more than 2 Gigawatt (2GW+) data centre and the expansion of its AI team.
Mark Zuckerberg, who termed 2025 as “a defining year for AI,” said in a Facebook post that the upcoming data centre is so large that “it would cover a significant part of Manhattan.” Zuckerberg also shows great faith in Meta’s Llama 4 large-language model. The billionaire said that this will become the leading state-of-the-art AI model. Notably, Llama 4 is set to release in early 2025. It is expected to be more capable than its predecessor, Llama 3, in terms of reasoning, adaptability, and performance.
Mark Zuckerberg further stated that the company would build an AI engineer to progressively contribute more code to its research and development projects.
Meanwhile, the company plans to end the year 2025 with more than 1.3 million graphics processing units. Notably, Meta is already one of the largest buyers of Nvidia’s AI chips, such as the H100. Clearly, Meta’s planned capital expenditures for 2025 in the AI sector are significantly higher than what analysts had expected. Zuckerberg’s projection of $60–65 billion is almost 64% higher than Meta’s expected spending of $38.4 billion in 2024.
Interestingly, the company is expecting a rapid rise in its AI assistant—Meta AI’s user base. Meta AI is currently available across all major social media platforms owned by the company, including WhatsApp, Facebook, Messenger, and Instagram. It also has a dedicated meta.ai portal. Zuckerberg believes that by the end of this year (2025), Meta AI will reach more than 1 billion users. If this happens, it would be a great achievement for the company, considering it reached 600 million monthly active users in 2024.
Meta’s latest announcement comes at a time when leading tech firms are continuously unveiling heavy investments in AI tech and infrastructure. In fact, a fresh example comes with the ‘Stargate Project’ announced by none other than US President Donald Trump. It is a $500 billion joint venture created by OpenAI, SoftBank, Oracle, and MGX. Under this project, the fund will invest to strengthen AI infrastructure in the United States by 2029, as rapid AI advancements are raising demand for the same.
Earlier this month, OpenAI’s prominent backer and tech giant Microsoft revealed massive investment plans to create AI data centers. The company announced it will invest around $80 billion in fiscal 2025 to construct data centers that are capable of supporting AI workloads. According to estimates, the global AI market was valued at around $233.46 billion in 2024. But by 2032, the market is expected to reach $1,771.62 billion.
However, the AI race is not as smooth as it seems for Meta. The company – which recently witnessed major policy overhauls to align with the new regime in the US – is facing some legal trouble related to its AI efforts. Meta faced a copyright infringement lawsuit back in 2023 by a group of authors. But now, in the amended lawsuit, the group of plaintiffs alleges that Meta used pirated content (including copyrighted books and articles) to train its Llama AI models with permission from its CEO, Mark Zuckerberg.
