Meta held its first AI developer conference Llamacon on Tuesday at its headquarters in Menlo Park, California. The company announced the launch of the consumer-facing Meta AI Chatbot app that will compete with Chatgpt, as well as a developer-facing API for accessing Llama models in the cloud.

Both distributions are designed to expand adoption of the company's open Llama AI model, but that goal may be a secondary to Meta's real motivation: defeating Openai. Meta's AI ambitions are sprawling and is fueling a thriving open AI ecosystem, pasting it into a "closed" AI provider like Openai, which lags its model behind the service.

Meta's AI Chatbot app almost feels like a preemptive move on the Openai rumored social network. It has a social feed where users can share their AI chats and provide personalized responses based on users’ meta-app activity.

As for the Llama API, this is a challenge for the OpenAI API business. The Llama API is designed to enable developers to build applications that connect to Llama models in the cloud using only one line of code. It eliminates the need to rely on third-party cloud providers to run Llama models and allows Meta to provide AI developers with more comprehensive tools.

Like many AI companies, Meta believes Openai is the highest competitor. In the case against META, court documents show that executives of the company previously defeated Openai's GPT-4, a former state-of-the-art model. Proprietary AI model providers like OpenAI have long been at the heart of Meta AI strategy. In a July 2024 letter, FW CEO Mark Zuckerberg tried to compare Meta with a company like Openai and wrote: "Selling access to AI models is not a business model (Meta).

Several AI researchers who spoke with TechCrunch before Llamacon hope Meta will release competitive AI inference models, such as Openai's O3-Mini. The company did not end up doing so. But for Meta, it's not about winning AI competitions.

In a stage conversation with Databricks CEO Ali Ghodsi, Zuckerberg said he saw any AI labs that make their models publicly available, including DeepSeek and Alibaba's Qwen, as a fight with closed model providers.

"So the part of the value around open source is that you can mix and match. So if other models (like DeepSeek) are better - or if Qwen is good at some way, as a developer, you can get the best smart parts from different models and produce exactly what you need," Zuckerberg said. “It’s part of the quality of open source basically through all the closed source (model) … (i) feels like an unstoppable force.”

In addition to stopping the growth of Openai, Meta may also try to push its open model to meet regulators. The EU AI Act grants special privileges to companies that distribute “free and open source” AI systems. Meta often claims that its Llama models are "open source" despite whether they meet the necessary standards.

Regardless of the reason, Meta seems content to launch AI, thus enhancing the open model ecosystem and limiting the growth of OpenAI – sometimes at the expense of failing to provide cutting-edge models themselves.