Meta commissions a Linux Foundation study to quantify open source AI's economics

What the study actually claims

On May 21, 2025, Meta published a report from Linux Foundation Research titled "The Economic and Workforce Impacts of Open Source AI," which it commissioned. The headline figures are specific: two-thirds of surveyed organizations believe open source AI is cheaper to deploy than proprietary models, and nearly half cite cost savings as a reason for choosing it.

The study also reports that 89% of organizations using AI in some form use open source AI, and it estimates companies would spend 3.5 times more if open source software did not exist. On the workforce side, it claims AI-related skills could raise a worker's wages by up to 20%.

The findings in this report make it clear: open source AI is a catalyst for economic growth and opportunity. As adoption scales across sectors, we're seeing measurable cost savings, increased productivity and rising demand for AI-related skills that can boost wages and career prospects.Montana Labs

The small-business claim is the interesting one

The most concrete and testable finding is that smaller businesses adopt open source AI at a higher rate than larger ones. Meta connects this to its own Llama Impact Grants, positioning small and medium firms as the primary beneficiaries of low-cost, customizable models.

This is a plausible dynamic. A team without a procurement budget or a vendor contract can pull down Llama weights and run them without per-token licensing. The friction that a large enterprise absorbs as a line item is a genuine blocker for a smaller shop. If the adoption gap is real, it says something practical about where open weights lower the barrier most.

Read the sector projections as ceilings, not forecasts

The report leans on McKinsey figures to size the opportunity: $170–$290 billion for advanced manufacturing and $150–$260 billion in value for global healthcare once AI is applied at scale across business functions. These are attributed to AI broadly, not open source AI specifically, and the manufacturing and healthcare examples describe potential integrations rather than measured deployments.

The healthcare framing — free, flexible tools aiding diagnostics in resource-limited hospitals — is the strongest case for why openness matters beyond price. But the dollar figures come from a different analysis about AI in general, so they establish an upper bound of the market, not a return anyone has booked from Llama.

The conflict of interest is the context, not a footnote

Meta commissioned this study, and Meta builds Llama. The report's closing line — that open AI models like Meta's Llama are becoming the industry standard — makes the commercial interest explicit. That does not make the survey data wrong, but it means the framing was chosen by the party that benefits from the conclusion.

For a team deciding between open and proprietary models, the useful signal here is the cost-of-deployment perception among peers and the adoption skew toward smaller organizations. The signal to discount is any single aggregate dollar figure, especially those borrowed from unrelated AI-wide projections. Treat the report as a well-sourced argument for open weights, not as neutral measurement of Llama's economic footprint.