Meta AI chief Alexandr Wang is facing internal and competitive challenges one year after the company's $14.3 billion AI investment [1].

These struggles place the company's aggressive pivot to artificial intelligence under scrutiny as Meta trails behind rivals including OpenAI, Anthropic, and Google. The company must now justify its massive spending amid a 19% decline in stock value [2].

Central to the friction is Muse Spark, Meta's first proprietary AI model. While the model represents a technical milestone, reports indicate that developers are largely ignoring Muse Spark [3]. This lack of adoption undermines the goal of creating a robust ecosystem for Meta's AI tools.

Internal stability has also deteriorated. High-profile AI hires have clashed with veteran staff, leading to layoffs and low morale [4]. Yann LeCun, a prominent AI researcher at Meta, has criticized the leadership of the division. "He is inexperienced," LeCun said [5].

Wang has pushed back against the idea that the company's talent pool is motivated solely by compensation. "It's unfair to say Meta's AI researchers are just there for the money," Wang said [6].

Despite these internal headwinds, some external indicators suggest a level of confidence in Meta-backed technology. The Pentagon recently awarded a $500 million contract to Scale AI, a firm backed by Meta [7]. This contract is five times larger than the deal awarded last year [7].

However, the disparity between government contracts and developer adoption creates a complex narrative for the company. Meta continues to struggle with morale issues, and LeCun has predicted further departures of key personnel [4].

“"He is inexperienced."”

The tension at Meta highlights the difficulty of integrating high-cost, external AI leadership into a legacy corporate culture. While the Pentagon contract suggests that Meta's technical infrastructure remains valuable to institutional clients, the failure of Muse Spark to gain developer traction indicates a gap between corporate capability and market utility. If Meta cannot stabilize its internal morale and increase model adoption, it risks a prolonged period of underperformance relative to its AI competitors.