Open-Source AI Models and Government Software Drive Funding and Adoption

Open-source and open‑weight AI models are distinguished by the extent of code, training data, and model weights made publicly available. The Open Source Initiative defines a fully open AI model as including both the trained weights and the underlying data, a standard not yet met by many recent releases such as Meta’s LLaMa, whose licence restricts commercial reuse. Meanwhile, venture capital is flowing into companies that specialise in scaling open‑weight infrastructure: Fireworks secured a $1.505 billion Series D round at a $17.5 billion valuation, Together AI raised $800 million, and Baseten attracted roughly $1.5 billion, despite not building their own foundation models.

In Canada, the government is expanding its use of open‑source software to accelerate public services. A senior developer at the Canadian Digital Service cites the GC Notify platform, which has sent close to 63 million notifications since its launch in November 2019. The service, originally built by the United Kingdom’s Government Digital Service, was adapted for bilingual delivery and further developed in an open manner so other agencies can reuse its components. The approach reduces duplication, speeds delivery, and has drawn attention from legislators, as the United States Senate considers a bipartisan bill to strengthen open‑source software adoption across federal agencies.