Meta raised its 2025 capital expenditure outlook to $64 billion to $72 billion to support its AI buildout and "expected cost of infrastructure hardware."
CEO Mark Zuckerberg quipped about Meta's spending on AI data centers during its LlamaCon conference this week. Meta had projected capital expenditures of $60 billion to $65 billion for 2025.
Going into the report, there were concerns about AI data center spending.
- A nuanced view of AI data center buildout concerns
- AI infrastructure becomes one big leveraged bet
- Meta launches Llama API, Meta AI app
- Meta plans to end 2025 with 1.3 million GPUs and $60B to $65B spent on AI
Meta can double down on its AI spending due to strong results in the first quarter. The company reported first quarter earnings of $16.64 billion, or $6.43 a share, on revenue of $42.31 billion, up 16% from a year ago.
Wall Street was looking for first quarter earnings of $5.28 a share on revenue of $41.4 billion.
Zuckerberg said the first quarter was a strong start and Meta AI was closing in on nearly 1 billion monthly active users. Meta said daily active people across its properties--Facebook, Instagram and WhatsApp--was 3.43 billion. and average price per ad was up 10% from a year ago.
As for the outlook, Meta projected second quarter revenue of $42.5 billion to $45.5 billion compared to estimates of $44.03 billion. Meta also said that it continues to "monitor an active regulatory landscape, including legal and regulatory headwinds in the EU and the U.S."
Speaking on the earnings conference call, Zuckerberg said:
"The major theme right now, of course, is how AI is transforming everything we do, and as we continue to increase our investments and focus more."
AI is driving new experiences, ad optimization, business messaging and building out Meta AI and AI devices, he said.
Meta is also using AI to train an ads ranking model that's twice as efficient improving ad efficiency. The generative ad recommendation model is trained on thousands of GPUs.
Regarding investments in AI infrastructure, Zuckerberg said the company continues to invest but is also optimizing its infrastructure and tweaking models to improve returns. He also added that it's important that Meta develops Llama models so it can control its own destiny.
The other reality is that Meta still can't meet the company's demand for compute resources.
Holger Mueller, an analyst at Constellation Research, said:
"Meta needs to ramp up it's capex as it needs to remain competitive with Google for the demand signals that matter to advertising. With Gemini 2.5 Google is taking share of searches and AI use of consumers that helps it's ad business. The question is of course how long can Meta - running Nvidia - can compete with Google running on TPUs - compete on the TCO side."
