Meta to spend up to $65 billion for big AI push in 2025

Meta CEO Mark Zuckerberg

Facebook parent Meta plans to invest $60 billion to $65 billion in capital expenditure in 2025 to bolster its artificial intelligence (AI) efforts, chief executive Mark Zuckerberg said on January 24, as the AI arms race heats up among tech giants.

This marks a significant increase from the company's planned $38 billion to $40 billion in capital expenditures for 2024.

Meta's announcement also comes on heels of Microsoft announcing plans to invest $80 billion in AI infrastructure in 2025. Additionally, US President Donald Trump recently unveiled a $500 billion AI infrastructure initiative, Project Stargate, with OpenAI, Oracle, Japan's SoftBank, and Abu Dhabi's MGX as partners.

"This is a massive effort, and over the coming years it will drive our core products and business, unlock historic innovation, and extend American technology leadership" Zuckerberg said in a post on Threads.

In October 2024, Meta had stated that it anticipates a "significant acceleration" in infrastructure expenses in 2025 due to expanded investments in servers, data centers, and other infrastructure. Zuckerberg had noted that their AI efforts demand 'serious infrastructure' and anticipates the company to continue making significant investments in this space.

In a post today, the Meta chief elaborated some of the investments the company is making in this area.

The company is building 2GW+ data center that is "so large it would cover a significant part of Manhattan," he said.

"We'll bring online ~1GW of compute in 2025 and we'll end the year with more than 1.3 million GPUs," Zuckerberg added. The social networking giant also plans to significantly expand its AI teams during the year.

Meta will also build an AI engineer that will start contributing increasing amounts of code to the firm's research and development efforts, he said.

"This will be a defining year for AI. In 2025, I expect Meta AI will be the leading assistant serving more than 1 billion people, Llama 4 will become the leading state of the art model" Zuckerberg said.

Llama 4, the next major version of the company's open-source AI model, is expected to launch in early 2025, beginning with smaller AI models, Zuckerberg said in October 2024.

The AI model will be a major upgrade over Llama 3 across several areas, including new modalities, capabilities, stronger reasoning, and much faster performance, he said.