Open-source AI emerges as middle-power answer to Big Tech control

Key takeaway

• Middle-power nations and the private sector are rallying around open-source AI as an alternative to closed models dominated by a handful of large corporations.

• Canada has prioritized open-source in its national AI strategy, while the EU has placed open source at the center of its technological sovereignty plans, with countries like Germany and Japan following suit.

• Open-source AI models surged from roughly 1–2% of developer token volume in late 2024 to nearly 30% by mid-2025, signaling strong market traction and reflecting efforts to ensure AI is transparent, accountable, and governed in the public interest rather than by a small number of actors.

3 Key Points

FAQ

What do Canada and the EU plan to do with open-source AI?

Canada has prioritized the development of open-source AI technology in its national AI strategy. The EU has placed open source at the centre of its technological sovereignty, with commitments to support open-source alternatives throughout the AI stack, backing startups, and developing new government procurement guidelines that favor open-source innovation.

How much have developers shifted toward open-source AI models?

Open-source models grew from roughly 1–2% of token volume in late 2024 to nearly 30% by mid-2025, according to a recent a16z and OpenRouter study.

Why is open-source AI seen as important for control and safety?

Unlike AI technology owned and controlled by a few large corporations, open-source AI is available to everyone, meaning governments and companies can own the infrastructure they build on. Additionally, anyone can look under the hood, and that transparency is what can make AI safe and accountable by design.