Meta's AI spending spree is Wall Street's focus in second-quarter earnings

Meta's Llama 4 AI model, a rushed attempt to mimic Chinese startup DeepSeek's architecture, disappointed developers and failed to significantly outperform rivals, prompting CEO Mark Zuckerberg to initiate a multi-billion dollar overhaul of the company's AI strategy. This includes a $14.3 billion investment in Scale AI, the formation of Meta Superintelligence Labs under newly appointed Chief AI Officer Alexandr Wang, and aggressive talent acquisition from competitors. While Meta's core ad business remains robust, Q2 revenue growth slowed to 15%, and analysts project increased CapEx and OpEx due to these AI investments. Despite internal debate over its open-source AI strategy, investors are largely receptive to Meta's intensified AI spending, viewing it as crucial for long-term competitive positioning in the high-stakes AI market.

Analysis

Meta Platforms is undergoing a significant and costly overhaul of its AI strategy following a material misstep with its Llama 4 model. The company's attempt to replicate the Mixture-of-Experts (MoE) architecture from Chinese startup DeepSeek resulted in a product that disappointed developers, eroding the goodwill established by its successful predecessor, Llama 3. This failure has triggered a multi-billion dollar course correction spearheaded by CEO Mark Zuckerberg, involving a $14.3 billion investment into Scale AI and the establishment of a new 'Meta Superintelligence Labs' led by high-profile hires from competitors. This strategic pivot coincides with notable financial headwinds, including a projected slowdown in Q2 revenue growth to 15%—its slowest rate since early 2023—and the potential for 2025 expenses to exceed the low end of the $113-$118 billion guidance. Internally, there is now a critical debate over the future of Meta's open-source AI strategy, with executives considering a shift towards a more powerful proprietary model. While the company's core advertising business remains strong, and investors seem more receptive to this AI spending than the prior metaverse push, the negative ticker sentiment (-0.3) reflects significant execution risk and uncertainty surrounding the return on these massive new investments.

Key Decisions for Investors

• Investors should closely monitor upcoming earnings calls for revised expense guidance and specific milestones from the new Meta Superintelligence Labs, given the significant near-term execution risk and cost pressures from the recent AI strategy overhaul.
• The potential pivot from an open-source to a proprietary AI model is a critical strategic consideration; any definitive change could alter the long-term investment thesis and should be scrutinized for its impact on the developer ecosystem.
• While slowing revenue growth and rising capital expenditures present short-term headwinds, investors should weigh these against the strength of the core ad business and the necessity of the current AI spending to maintain competitive parity.
• Future investment decisions should be contingent on the new AI unit's ability to deliver a competitive model that regains developer trust, as the Llama 4 stumble proves that increased spending alone does not guarantee market leadership.