Meta no longer prolonging the life of datacentre servers

Meta, the parent company of Facebook, Instagram, and WhatsApp, has announced a 19% rise in revenue to $40.6 billion for the third quarter of 2024. This growth is attributed, in part, to the advancements in artificial intelligence (AI) as highlighted by founder and CEO Mark Zuckerberg.

Zuckerberg stated, “We had a successful quarter driven by AI advancements across our apps and business. We are also seeing significant progress with Meta AI, Llama adoption, and AI-powered glasses.”

During the earnings call discussing the quarterly results, Chad Heaton, VP of Finance at Meta Platforms, addressed inquiries about the return on investment from Meta’s capital expenditure (CapEx). With reported costs and expenses up by 15% and $9.2 billion in CapEx, Heaton discussed the company’s expansion of datacenters and servers to support the AI and non-AI services offered by Meta’s various businesses.

Heaton mentioned, “We are in the process of finalizing our infrastructure plan for 2025. We anticipate an increase in total infrastructure spending for non-Gen AI, GenAI, and core AI in the upcoming year.”

The company intends to primarily invest in servers, with additional spending on datacenters and network equipment to enhance Meta’s IT infrastructure. Heaton highlighted plans for server investments to bolster user engagement in non-AI capacities and to upgrade existing servers.

Furthermore, Meta plans to initiate the construction phase of its strategy to develop smaller, high-density datacenters. Heaton emphasized, “We are expanding our networks to accommodate the growth in GenAI and core AI-related traffic in 2025, with investments in fiber optics to handle future cross-region training traffic.”

When asked about extending the lifespan of datacenter servers to reduce costs, Meta’s CFO Susan Li mentioned, “We do not currently have plans to extend the useful lives of our servers. Our focus is on utilizing the latest generation of servers to maximize the capacity in our datacenters.”

Aside from hardware, Li highlighted the company’s efforts in optimizing existing recommendation systems to drive improvements and efficiency.

Discussing recent enhancements on Instagram, she explained, “We completed efficiency projects that consolidated existing recommendation systems, allowing us to concentrate on fewer models for future improvements.”

Li emphasized the importance of driving higher engineering efficiency through these optimizations.

Regarding return on investment, she mentioned, “Our focus is on enhancing relevance as we scale our models for content recommendation.”

This involves increasing model complexity to leverage more compute power for better analysis of existing data. Meta also plans to expand the volume of data sets used in AI to enable machine learning algorithms to learn from a wider range of data.