IE 11 is not supported. For an optimal experience visit our site on another browser.

Meta's big, expensive AI bet hinges on giving its models away for free

Meta recently announced the rollout of Llama 3, the open-code large language model that underpins the Meta AI assistant tool widely accessible on its flagship social media platforms.
Mark Zuckerberg
Meta CEO Mark Zuckerberg in Menlo Park, Calif., in September.Josh Edelson / AFP via Getty Imagesfile

Meta is spending big on artificial intelligence — only to offer its latest advanced AI models to the public for free.

The tech giant recently announced the rollout of Llama 3, the open-code large language model that underpins the Meta AI assistant tool now widely accessible on its flagship social media platforms. Meta aimed to “build the best open models that are on par with the best proprietary models available today,” it wrote in a blog post.

The model quickly drew acclaim from technologists — but not from investors.

Meta stock plummeted Wednesday after CEO Mark Zuckerberg touted a slew of money-losing projects on its first-quarter earnings call. Zuckerberg expressed confidence in investing even further in advanced open-source AI models.

“I’m very pleased with how Llama 3 has come together so far,” he said on the earnings call, adding, “I expect our models are just going to improve further from open-source contributions.”

“Open source” is generally used to describe software that is made available to the public, which can use the code and improve it or build on it. Several rapidly growing AI startups, such as Mistral AI and Hugging Face, have found success offering open-source models and tools.

In AI, models — such as large language models and foundational models — are the complex pieces of software that use algorithms, trained on relevant data, to complete tasks that include recognizing patterns and making predictions or decisions. OpenAI, Google, Anthropic and many more companies are constantly developing, training and tweaking their models seeking to create ever more capable AI programs.

So what’s the catch? Why would Meta spend so much time and money developing AI models only to make them open source?

Alex Ratner, CEO of the data-focused company Snorkel AI and an assistant professor of computer science and engineering at the University of Washington, wrote in an email that Llama 3 is a “much bigger step change than many people realize.” He said he suspects it will benefit AI developers en masse while also enabling Meta to reduce costs, attract talent and boost potential revenue.

When a business or an organization wants to integrate AI technologies into its operations, it must use its own internal data to properly tune those AI models. But using closed-source models means giving that data — often the most valuable company asset — away to an AI model provider who ultimately owns the final model.

Using a model like Llama 3, however, gives those enterprises a way to maintain ownership of their own data, as well as the output of their work. That makes using open-source models much more appealing to developers and data scientists, he wrote, possibly giving Meta better chances to attract top talent in the AI space.

“Enterprise AI is becoming less about the biggest, most powerful generalist model, and more about the specialist trained on your data,” Ratner wrote, “and Meta has positioned Llama 3 to potentially become the epicenter of that.”

Others in the tech community have offered similar assessments. Bindu Reddy, CEO of the generative AI startup Abacus.AI, said recently on X that Meta’s move with Llama was “a strategic masterstroke and serves its business interests.”

Aaron Levie, CEO of the cloud computing company Box, said on X: “Incredible to have Meta out here pushing open source AI forward in such a major way. An amazing moment to be building software when you have $100s of billions of dollars from the major tech players going into advanced R&D and infra that directly benefit you as a developer.”

Unlike many of its fellow tech giants, Meta has a history of investing in open-source projects, from the machine learning library PyTorch to the JavaScript library React to the data query language GraphQL, among others. Its earlier iterations of Llama were also advertised as open source.

But there are limits to Llama’s openness.

While Meta allows most developers to use Llama 3 commercially, those who seek to use it in products with more than 700 million monthly active users — essentially, just major tech platforms such as Google, TikTok and Snapchat — must request licenses from Meta. It also has not disclosed the data used to train Llama 2 and 3, which makes it hard to gauge potential bias.

And the company’s license for Llama 3 is not approved by the Open Source Initiative, a widely recognized nonprofit authority that sets rules defining open-source software, which makes labeling Meta’s LLM “open source” somewhat controversial in tech spaces.

Still, Zuckerberg laid out in another earnings call this year that Meta has strategic interests in positioning itself as a leader in open-source software infrastructure. Not only is open-source software a more popular option among developers and researchers, he said, but such software is also more likely to become an industry standard.

The advantage with closed-source AI models — like OpenAI’s GPT, Anthropic’s Claude and Google’s Gemini — is that the AI providers behind the LLMs retain more control over their proprietary algorithms and technologies, making it harder for competitors to replicate them. But Zuckerberg has said he is not worried.

“The short version is that open sourcing improves our models,” Zuckerberg said on the call. “And because there’s still significant work to turn our models into products and because there will be other open-source models available anyway, we find there are mostly advantages to being the open-source leader, and it doesn’t remove differentiation from our products much anyway.”