Meta is exploring the integration of artificial intelligence (AI) models from Google and OpenAI into its applications. Executives have reportedly considered incorporating Google's Gemini model into Meta AI, a model that has already appeared in chatbots across its platforms.
This potential collaboration suggests that the company, led by Mark Zuckerberg, may temporarily rely on rival technologies to enhance its AI capabilities for users. At the same time, Meta continues to develop its own advanced systems. According to reports, the company’s AI leadership views this strategy as a short-term move rather than a permanent shift. Meta is currently working on Llama 5, the latest version of its in-house AI model, which is still undergoing development and training. The company believes that once Llama 5 is ready, it will be capable of competing directly with models like Google's Gemini and OpenAI's offerings. Until then, utilizing external systems could help the social media giant maintain the competitiveness and appeal of its AI products.
It’s worth noting that Meta has long advocated a hybrid AI strategy that includes building its own systems, collaborating with other companies, and supporting open-source initiatives. While advancing its own models, the company is actively exploring beneficial partnerships. In fact, Meta has already integrated external technologies into some of its internal tools. For example, its engineers rely on AI models from Anthropic to assist with coding tasks via internal tools.
Recently, Meta also entered into a collaboration with Midjourney, the AI lab known for its image generation tools. This partnership goes beyond a simple licensing agreement, as Meta plans to use Midjourney’s technology and work directly with its researchers to develop more powerful image and video AI models. For Meta, this collaboration offers a faster path to improving the quality of its AI-generated visual content, as it seeks to catch up with tools like Google’s Veo 3 for video and OpenAI’s Sora.
Despite these collaborations—even with competitors—Meta remains heavily invested in its own AI research and development. The company has raised its annual capital expenditure forecast to between $66 billion and $72 billion. Earlier this year, it hired Alexandr Wang (former CEO of Scale AI) and Nat Friedman (former CEO of GitHub) to co-lead its new initiative, the Meta Superintelligence Lab. However, its AI R&D efforts still face challenges and controversies. Its latest large language model, Llama 4, did not achieve the expected level of adoption. Additionally, its AI chatbots have faced criticism after reports surfaced of them engaging in sexually explicit conversations with users, including minors. In response to these challenges, the company appears to be revising its recently adopted AI strategy. Recent reports indicate that Meta is reshaping its AI operations for the fourth time in just six months. Under the new plan, Meta will split its AI division.
