Mozilla's Inaugural 'State of Open Source AI' Report Is Here

3 hours ago

Open source AI models have nearly closed the performance gap with proprietary systems like ChatGPT and Claude, with only a 3% difference, while costs have dropped up to 50x in three years.
Open models power about one-third of real-world AI usage but capture only 4% of revenue, indicating value isn't flowing back into the open ecosystem.
China and East Asia lead in open source AI adoption at 89%, treating it as a national strategy, while many countries are enacting new AI strategies and restricting foreign data processing.
79% of developers use open models, but only 51% have deployed them in production, highlighting a gap in infrastructure and tooling rather than model quality.
The key competitive layer is now the 'agentic harness'—software controlling what AI can see, remember, and do—with few guardrails and high user consent fatigue.
Mozilla's report emphasizes the need to invest in infrastructure, tooling, and governance to ensure open AI can scale in the public interest.

Open-weight AI is being successfully adopted in diverse, real-world applications globally, from Māori language models to agricultural diagnostics, emphasizing ownership and control.
A Mozilla/SlashData survey shows 79% of developers using AI choose open models, though only 51% of open-model teams reach production compared to 63% for closed, due to operational tooling and trust gaps.
Open models have become a commercial market at a multi-hundred-billion-dollar scale, with companies like Databricks, Mistral, and DeepSeek achieving significant revenue and valuation growth.

The author debated with Richard Stallman in the 1980s about open source software, initially holding the conventional view that proprietary control was necessary for software advancement, but eventually was convinced by Stallman that software as a body of knowledge grows stronger when shared.
Stallman's free software movement evolved into open source, leading to key projects like GCC and GNU/Linux, which now power much of the internet, demonstrating that transparency and community collaboration enhance security and innovation rather than hinder it.
Open source software has accelerated technological progress by creating a shared knowledge base that educates engineers and supports a delicate ecosystem of private companies, universities, and volunteers, with many successful companies built on open core models.

US AI strategy relies on proprietary, locked-down models, while China adopts an open-weights approach.
AI models lack strong moats; switching between them is easy via APIs, making enterprise services the real differentiator.
China's open models exploit US export controls by turning compute disadvantages into distribution advantages and commoditizing the AI layer.

DeepSeek's open-source R1 model released in January 2025, which rivals top US AIs and is free, caused a $1 trillion drop in US tech stocks and prompted US lawmakers to propose banning it on government devices.
Z.ai's GLM-5.2 release in contrast did not cause similar panic, indicating a shift in the US-China AI arms race where China now catches up, similar to other technologies like smartphones and electric cars.
Chinese AI models are typically open-source and free for local use, while US models are often cloud-based, fee-charging, and closed, leading to differences in accessibility and innovation strategies.

Open source software is a critical foundation for the global economy, enabling accessibility and community oversight in key sectors.
Cybersecurity is a top beneficiary of open source, and the Open Secure AI Alliance aims to use open technologies to fix vulnerabilities.
Open AI models are essential for democratizing defense, increasing transparency, and providing customizable controls without single points of failure.

Open-weight AI is being successfully adopted in diverse, real-world applications globally, from Māori language models to agricultural diagnostics, emphasizing ownership and control.
A Mozilla/SlashData survey shows 79% of developers using AI choose open models, though only 51% of open-model teams reach production compared to 63% for closed, due to operational tooling and trust gaps.
Open models have become a commercial market at a multi-hundred-billion-dollar scale, with companies like Databricks, Mistral, and DeepSeek achieving significant revenue and valuation growth.

The author debated with Richard Stallman in the 1980s about open source software, initially holding the conventional view that proprietary control was necessary for software advancement, but eventually was convinced by Stallman that software as a body of knowledge grows stronger when shared.
Stallman's free software movement evolved into open source, leading to key projects like GCC and GNU/Linux, which now power much of the internet, demonstrating that transparency and community collaboration enhance security and innovation rather than hinder it.
Open source software has accelerated technological progress by creating a shared knowledge base that educates engineers and supports a delicate ecosystem of private companies, universities, and volunteers, with many successful companies built on open core models.

US AI strategy relies on proprietary, locked-down models, while China adopts an open-weights approach.
AI models lack strong moats; switching between them is easy via APIs, making enterprise services the real differentiator.
China's open models exploit US export controls by turning compute disadvantages into distribution advantages and commoditizing the AI layer.

DeepSeek's open-source R1 model released in January 2025, which rivals top US AIs and is free, caused a $1 trillion drop in US tech stocks and prompted US lawmakers to propose banning it on government devices.
Z.ai's GLM-5.2 release in contrast did not cause similar panic, indicating a shift in the US-China AI arms race where China now catches up, similar to other technologies like smartphones and electric cars.
Chinese AI models are typically open-source and free for local use, while US models are often cloud-based, fee-charging, and closed, leading to differences in accessibility and innovation strategies.

Open source software is a critical foundation for the global economy, enabling accessibility and community oversight in key sectors.
Cybersecurity is a top beneficiary of open source, and the Open Secure AI Alliance aims to use open technologies to fix vulnerabilities.
Open AI models are essential for democratizing defense, increasing transparency, and providing customizable controls without single points of failure.

Hasty Briefsbeta