Meta’s Absence from New AI Standards Group Sparks Industry Shift
204Meta has taken a different route in the evolving AI landscape. This week, The Linux Foundation announced the formation of a new group called the Agentic AI Foundation (AAIF). Its goal is to create shared tools, standards, and foster community-driven innovation for developing AI agents. Many big tech names, including AWS, OpenAI, Google, Microsoft, IBM, and Cisco, are part of the group. But Meta is noticeably missing from the list. This move hints at Meta’s plans to stay independent and focus on its own AI projects.
Meta’s Focus on Proprietary AI Models
According to a recent report, Meta is working on a new proprietary AI model codenamed Avocado. Unlike open-source projects, this model is designed to generate revenue for the company. Industry experts say Meta prefers to keep control over its training data and how its models are governed. Brian Jackson, a research director, explained that Meta isn’t interested in sharing its training data, which it sees as a competitive advantage. Instead, the company wants to maintain control over how its models are used and integrated with other platforms.
Jackson noted that Meta’s approach is different from openly sharing AI weights or model parameters. Weights are the parts of a neural network that can be tweaked during training. Sharing these weights is a step toward open source, but Meta is holding onto its training data to protect its interests. The company seems to believe that open weights alone aren’t enough to define standards or distribute models in the way it intends. This stance positions Meta as a competitor that values control and monetization over open collaboration.
The Industry’s Shift Toward Monetization
Developing advanced open-source AI models is becoming more expensive. At the recent AWS re:Invent conference, AWS CEO Matt Garman shared insights about open weights models. He pointed out that open source software benefits from community contributions, but open weights models are mostly contributed to by the provider. This makes them costly to develop and maintain, and many companies are realizing they need to charge for access to sustain their investments.
Jackson added that Meta’s lack of a clear revenue model for its open weights models indicates a shift toward monetization. Some speculate that Meta’s strategy was to commoditize large language models (LLMs) and weaken competitors’ businesses. As these models grow larger and more complex, the costs to train and support them also increase. To stay competitive, Meta will likely need to monetize its AI efforts by offering access through gated APIs and charging per token processed.
Expert opinions suggest that Meta’s move to focus on proprietary models and potential monetization reflects a broader industry trend. As AI models become more resource-intensive, companies are looking for ways to turn their investments into revenue. Whether through APIs, licensing, or other services, the push toward commercialization is reshaping how AI development is approached across tech giants.
Overall, Meta’s absence from the new standards group and its focus on proprietary AI models highlight a key industry shift. While open-source communities continue to push for shared standards, major players like Meta are choosing control and monetization. This could influence how AI tools and models evolve in the coming years, impacting both competition and collaboration across the tech industry.
What do you think?
It is nice to know your opinion. Leave a comment.