Meta’s effort to compete in artificial intelligence hit a major setback when Google cut off the company’s access to its advanced Gemini AI platform earlier this year, according to multiple reports. This move significantly disrupted Meta’s AI initiatives, including crucial projects like data moderation and customer service automation, underscoring how dependent Meta has become on external AI technologies.
The decision to rely heavily on Google’s Gemini came after Meta found its own open-source Llama AI models underperforming. Sources revealed that Meta opted for Gemini because it delivered superior results, even though Google had to limit access due to high demand among clients. Meta’s uniquely large request volume for Gemini models contributed to Google’s resource constraints.
Amid these challenges, Meta’s AI division has faced internal struggles ranging from low employee morale and executive departures to rising public relations issues. Despite a massive financial commitment scheduled over the next two years to revamp AI capabilities, the company has had to reverse its initial encouragement for employees to aggressively use AI. Instead, staff are now being instructed to conserve AI usage due to rising costs.
Meta is working to develop its own AI alternatives and recently launched Muse Spark, a multimodal reasoning model. The company describes this as a foundational step toward reducing dependence on competitors’ technology. However, insiders indicate that even for Spark’s development, Meta had to seek Google’s help, highlighting ongoing difficulties in achieving AI independence.