Meta Unveils Llama for Startups Program to Propel AI Innovation

>Meta recently announced a great new program to increase innovation among new startups. First, they allowed these companies access to their Llama deep-learning AI models. The world’s most valuable company is making this announcement amid a fierce global race toward artificial intelligence. It’s up against formidable competition, such as DeepSeek, Google, and Alibaba’s Qwen. The program, which is known officially as Llama for Startups, features hands-on support from Meta’s Llama team and funding opportunities for eligible companies.

Meta’s announcement that it deployed its Llama 4 Maverick model to address the LM Arena benchmark is very timely. This homegrown but globally popular crowdsourced AI test is what got the company to such an impressive score. This success, though, came alongside serious allegations of impropriety — which Meta has since taken steps to remediate. Alongside the announcement of the Llama for Startups program, the company made a different version of the Maverick publicly available.

Meta announced in a blog post that U.S.-based businesses of any size are eligible to apply for the program, provided they satisfy certain requirements. Save the date— May 30 is the deadline! Eligible companies need to be incorporated and have raised under $10 million in funding. They need to be staffed by a team of developers dedicated to pushing out generative AI applications.

The Llama for Startups program was designed with the intention of alleviating the financial burdens associated with participation.

“Members may receive up to $6,000 per month for up to six months to help them offset the costs of building and enhancing their generative AI solutions.” – Meta

Meta’s startup experts will work one-on-one with startups and small businesses. Jointly, they will prototype some innovative, cutting-edge use cases of the Llama models to improve their operations.

The releases of Meta’s Llama models have proven to be highly consequential and quickly adopted, amassing over one billion downloads. In fact, several of the companies hosting these models have already signed revenue-sharing agreements with Meta. That’s an indication of the high demand for its AI technology. Furthermore, Meta recently introduced an API designed for customizing Llama releases, enabling developers to tailor the models to their specific needs.

That means capital expenditures, which include investments in data centers and servers, will total $60 billion to $80 billion come 2025—just for Meta. Their key focus area will be siting new data centers. This funding is a much-needed step to strengthen the infrastructure that will actually support its growing base of new AI projects.