Meta officially launched the Llama Startup Program on 21st May 2025, providing 50 selected ventures with access to the complete Llama 3.1 model family, including the 405 billion parameter Llama 3.1 405B model, alongside technical support and up to $150,000 in Meta AI cloud credits. The programme aims to foster AI solution development in critical areas such as healthcare, education, and climate change, whilst equipping participants with cutting-edge tools to create sophisticated AI applications.
The initiative is built on three key pillars: unlimited access to Llama models, dedicated engineering support from the Meta AI team, and free cloud computing resources for development and testing. Shashi Narayan, Lead Researcher at Meta AI, emphasised that the programme specifically targets startups developing socially beneficial AI products, with particular focus on four main areas identified by Meta: healthcare applications, educational tools, environmental solutions, and innovations aimed at reducing the digital divide. During the selection process, Meta pays special attention to applications from diverse geographical regions, and the programme is launching with a limited cohort of 50 startups in its first year.
The Llama Startup Program complements Meta's existing Llama innovation ecosystem, which includes the Llama Impact Accelerator programme that provides additional support for ventures focused on social impact. According to Meta's data, the current open-source version of Llama models has already attracted over 3 million developers worldwide, and the company plans a 50% growth in this figure by the end of 2025, clearly indicating the rapid development of the AI industry and Meta's strategic commitment to supporting AI innovation.
Sources:
1.

2.

3.
