Meta to defend AI spending ahead of Q2 earnings call

Meta is set to defend its major AI investments and recent strategic changes during its second-quarter earnings call on July 30, 2025.

The US-based social media giant has faced challenges after its latest Llama 4 AI model, which reportedly tried to mimic techniques from Chinese startup DeepSeek, received a lukewarm response from developers.

In response, CEO Mark Zuckerberg authorized billions in additional AI spending and announced a significant hiring spree, and Meta invested US$ 14.3 billion in Scale AI, according to people familiar with the matter.

Meta also said Alexandr Wang, CEO of Scale AI, along with other industry leaders like Nat Friedman and Daniel Gross, will help lead the new Meta Superintelligence Labs.

Meta’s shift to a mixture-of-experts (MoE) AI model with Llama 4 was intended to compete with rivals like OpenAI, though some developers prefer the previous Llama 3 model for its ease of integration.

Executives are now reassessing Meta’s open-source AI approach, with internal debate over whether to keep releasing models publicly or shift to proprietary technology, sources say.

Despite AI setbacks, Meta’s core advertising business remains strong.

🔗 Source: CNBC

🧠 Food for thought

1️⃣ Big tech’s AI infrastructure race reaches unprecedented scale

Meta’s AI hiring spree reflects a broader pattern of massive infrastructure investments across major tech companies.

Google has committed US$85 billion to AI and cloud infrastructure by 2025, while Amazon allocated $100 billion for AI capabilities in its cloud division1.

Meta’s projected spending of US$64-72 billion in 2025 puts it in the same league as these tech giants, showing how AI competition has escalated beyond software development into fundamental infrastructure buildout1.

This represents a dramatic shift from previous technology cycles, where companies could compete primarily through software innovation rather than requiring tens of billions in physical infrastructure investments.

The scale suggests that AI leadership may increasingly depend on which companies can sustain the highest capital expenditure levels, potentially creating barriers for smaller competitors who cannot match these investment levels.