Tariffs could cost Meta $8bn in extra datacentre costs
As the company boosts its investment in infrastructure for artificial intelligence (AI), Meta’s latest earnings call has painted a bleak picture for software developers.
The social media giant posted first-quarter 2025 revenue of $42.3bn, a 16% increase from the same quarter in 2024. Cost of revenue increased 14%, driven primarily by higher infrastructure costs and payments to partners, partially offset by a benefit from the previously announced extension of server useful lives.
When asked about a recent conversation about AI in coding he had with Microsoft CEO Satya Nadella at the LlamaCon conference, Meta CEO Mark Zuckerberg claimed the ability of AI to write code was reaching the level of a mid-level engineer. He added that by next year, the majority of the company’s research with AI will be achieved using AI agents.
Zuckerberg also predicted that in the next few years, every business will have an AI business agent for customer support and sales, just as they all have an email address and social media accounts.
The company is heavily pushing AI capabilities such as large language models (LLMs) to power its recommendation engines and serve adverts to users.
Chief financial officer Susan Li said: “A big focus of this work will be on developing increasingly efficient recommendation systems so that we can continue scaling up the complexity and compute used to train our models while avoiding diminishing returns. We’re finding that LLMs’ ability to understand a piece of content more deeply than traditional recommendation systems can help better identify what is interesting to someone about a piece of content, leading to better recommendations.”
Among the growth opportunities is AI devices, according to Zuckerberg. “This is increasingly how we’re thinking about our work on the next generation of computing platforms,” he said. “Glasses are the ideal form factor for both AI and the metaverse.”
With more than a billion people worldwide who wear glasses today, Zuckerberg anticipated that over the next five to 10 years, most eyewear will incorporate artificial intelligence.
To support the company’s AI efforts, he said Meta was accelerating efforts to bring datacentre capacity online more quickly this year, adding that there was also work underway to provide Meta with the flexibility to add capacity in the coming years.
Li said Meta would be spending more this year on both generative AI and IT infrastructure to support core business. “We expect the significant infrastructure footprint we’re building will not only help us meet the demands of our business in the near-term, but also provide us with an advantage in the quality and scale of AI services we can deliver,” she said.
Li added that Meta would continue to build datacentre capacity in a way that “grants us maximum flexibility in how and when we deploy it to ensure we have the agility to react to how the technology and industry develop in the coming years”.
She also said Meta was working to increase the efficiency of its datacentre workloads. “Many of the innovations coming out of our ranking work are focused on increasing the efficiency of our systems,” said Li. “This emphasis on efficiency is helping us deliver consistently strong returns from our core AI initiatives.”
The company stated that its forecast for capital expenditure would be increasing to between $65bn and $72bn, which represents an increase of between $4bn and $7bn compared with what it had previously forecast.
When asked if Meta’s higher capital expenditure on datacentre infrastructure was a result of the US administration’s new trade tariffs, Li said: “The higher costs we expect to incur for infrastructure hardware this year really comes from suppliers who source from countries around the world. And there’s just a lot of uncertainty around this, given the ongoing trade discussions.”
She added that Meta would be working on mitigations by optimising its supply chain. “Our outlook is really trying to reflect our best understanding of the potential impact this year across all of that uncertainty,” said Li.