Meta’s Strategic Pivot: From Open Source to a Closed AI Model

Meta US30303M1027

A fundamental shift is underway in Meta's approach to artificial intelligence. The social media giant is preparing to abandon its open-source philosophy, moving instead toward a proprietary and monetizable AI framework. This strategic reversal coincides with investor unease over the company's ballooning infrastructure investments.

Central to this change is an internal initiative, reportedly codenamed "Avocado." The project aims to launch a closed AI model by early 2026. Key aspects of this development include:
* Leveraging technology partially sourced from Chinese tech firms, including Alibaba.
* Maintaining Meta's full commercial control over the final product.
* Representing a direct response to perceived shortcomings of the open-source Llama 4 model.

This move signals a clear corporate realization: while open-source development drives innovation, it offers limited direct monetization pathways, especially when compared to the revenue models of rivals like OpenAI and Alphabet.

Analyst Outlook: Lowered Targets Amid Sustained Confidence

On December 11, investment bank Morgan Stanley adjusted its position on Meta, reducing its price target from $820 to $750. The firm maintained its Overweight rating but cited concerns over advertising revenue sustainability and uncertainty surrounding massive AI expenditures. The bank expressed particular skepticism regarding the Llama 4 open-source strategy.

Despite this reduction, the new target still implies significant upside from the current trading level of approximately $650. Analysts underscored that Meta remains one of the few corporations with the data assets, global reach, and financial capacity to convert long-term AI investments into substantial value.

Soaring Expenditures and Strategic Focus

Meta's capital investment plan continues to be a point of contention. For 2025, the company anticipates capital expenditures in the range of $66 to $72 billion. Management has further guided toward "significantly higher" spending in 2026, with some estimates surpassing $100 billion. This effectively channels nearly all of Meta's operating cash flow into building out its AI infrastructure.

In a related strategic tightening, Meta has paused its program to license its Horizon OS for third-party VR headsets. Previously announced partnerships with ASUS and Lenovo are now on hold, indicating a renewed focus on internal hardware priorities.

Collaborations and Long-Term Optimism

In a development that could increase flexibility and reduce costs, Reuters reports that Google is collaborating with Meta to enhance PyTorch's compatibility with Google's Tensor Processing Units (TPUs). This cooperation may lessen Meta's dependence on NVIDIA's CUDA ecosystem and lower AI inference costs as new features scale across Instagram, Facebook, and WhatsApp.

The broader Wall Street consensus remains largely optimistic. The average analyst price target sits near $819, suggesting roughly 26% potential upside. Rosenblatt Securities presents an even more bullish case, with a $1,117 price target, citing disciplined cost management in the Metaverse division and a strategic shift toward AI-powered wearables.