Meta has introduced the ‘Llama Startup Program’, a strategic initiative aimed at empowering early-stage companies in the United States to leverage its open-source Llama AI models. This initiative provides financial assistance and technical mentorship to help startups integrate Llama into their products.
Under this initiative, eligible startups can receive up to $6,000 per month for six months (totaling $36,000) to compensate for development costs. The program is open to US-incorporated companies that have raised less than $10 million in funding and employ at least one developer.
Along with funding, participants will also receive guidance from Meta’s engineering teams to explore advanced use cases of Llama that could benefit their startups. As per the company, applications for the first group in the program will close on May 30, 2025, at 6:00 pm PT. Notably, last month, the Mark Zuckerberg-led firm launched Llama 4 Scout (featuring 17 billion active parameters and using 16 experts) and Llama 4 Maverick (also with 17 billion active parameters but using 128 experts).
“Early-stage startups are agile and creative, making them uniquely positioned to accelerate high-impact innovation with Llama. In a recent Linux Foundation study, 94% of organizations say they’ve already adopted AI tools and models, and of those, 89% are using some form of open source technology (such as Llama) in their AI infrastructure. The Llama Startup Program is ideal for those building generative AI applications across a variety of industries, including technology and software, financial services, healthcare and life sciences, telecommunications, and retail and eCommerce,” the company said in a blog post.
The move comes at a time when, despite Llama’s impressive milestone of over a billion downloads, rival companies like OpenAI, Google, and DeepSeek (with their own AI models) are posing tough challenges. Notably, Meta’s development of Llama has not been without its setbacks. Recent reports suggest that the social media powerhouse has postponed the release of its flagship AI model, Llama 4 Behemoth, due to concerns over its performance and readiness. It was originally set to launch in April 2025, then delayed to June, and is now expected sometime in the fall (September-November) or later.
The development challenges have also caused frustration among Meta executives, prompting considerations for restructuring the AI product group responsible for Llama 4. Making things worse, 11 out of the 14 researchers who worked on the original Llama model have reportedly left the company. The company also faced criticism for allegedly manipulating benchmark results.
The struggle of Meta with its AI initiative is not new. For example, Meta’s AI chatbot has recently come under intense scrutiny following reports that it engaged in sexually explicit conversations with users, including minors. Additionally, the company is also facing a copyright infringement lawsuit, which alleges that Meta used pirated content to train its LLaMA AI models with the approval of its CEO, Mark Zuckerberg.
All this is happening when, at the start of this year, CEO Mark Zuckerberg termed 2025 “a defining year for AI” and announced the company’s plans to invest around $65 billion to boost its artificial intelligence infrastructure. In April 2025, the social media giant introduced a standalone AI assistant app powered by its latest large language model (LLM) – Llama 4.
