May 22nd.Meta Announcing the launch of a new program called "Llama for Startups," designed to encourageStartupsusing its Llama AI Models. The program provides eligible startups with "direct support" from the Meta Llama team and, in some cases, financial support.
According to Meta's official blog, any U.S.-incorporated startup with less than $10 million in funding (note: that's about Rs. 72.015 million at the current exchange rate), at least one developer, and that's building a generative AI application has until May 30 to apply to join the program.Selected members receive up to $6,000 per month in financial support,Up to six monthsto help offset the cost of building and enhancing generative AI solutions. meta also said its team of experts will work closely with startups to help them get off to a fast start and explore advanced use cases for Llama to help grow their business.
The launch of the Llama startup program comes at a time when Meta is trying to solidify its leadership position in the highly competitive open source modeling space. While Meta's Llama model has garnered more than a billion downloads to date, competitors including DeepSeek, Google and Alibaba's Qwen are threatening Meta's efforts to build a broad ecosystem of models.
Adding insult to injury, Llama has suffered several setbacks in the past few months. Last week the Wall Street Journal reported that Meta delayed the launch of its flagship AI model, Llama 4 Behemoth, over concerns that it had performed poorly in key benchmarks. In April, Meta also had to deal with allegations of cheating on the popular crowdsourced AI benchmark LM Arena. The company used a "conversationally optimized" version of the Llama 4 Maverick model to score high at LM Arena, but released another version of the model publicly.
That said, Meta still has big expectations for Llama and its broader generative AI portfolio. Last year, the company predicted that its generative AI offerings would generate $2 billion to $3 billion in revenue by 2025 and $460 billion to $1.4 trillion by 2035. meta has revenue-sharing agreements with a number of companies hosting its Llama models, and recently launched an API for customizing versions of Llama. Additionally, Meta AI, Meta's AI assistant (powered by Llama), may in the future display ads and launch a subscription service with additional features.
However, these products are expensive to develop. in 2024, Meta's budget for "generative AI" (GenAI) was over $900 million, and this year it could be over $1 billion, not including the cost of the infrastructure needed to run and train the models. meta has previously stated that it plans to spend $60 billion to $80 billion on capital expenditures in 2025, mainly on new data centers. Meta has previously stated that it plans to spend between $60 billion and $80 billion on capital expenditures by 2025, primarily on new data centers.