Arcee AI Raises $150 Millon to Build American Open Models That Enterprises Can Own
Any company that uses ChatGPT or Claude through an API doesn’t really own the model. They send data to someone else, pay per token, and cannot see how the model The post Arcee AI Raises $150 Millon to Build American Open Models That Enterprises Can Own appeared first on Ventureburn .
Arcee AI has secured $150 million in funding to develop American open models that can be owned by enterprises. These models would run inside companies' own data centers, addressing concerns about data privacy and control over AI systems. The Series B funding round was led by Vista Equity Partners, Cambium Capital, and Emergence Capital, with support from Microsoft’s venture arm M12, AI10 Ventures, Hitachi, IAG, P7, and Wipro.
The startup's valuation was boosted to over $1 billion, making it a new unicorn. Arcee's CEO McQuade invested $30 million of this capital in training new models, with the majority allocated to developing the Trinity family of models. These models, released in early 2026, are 400 billion parameter mixture of experts models consisting of 13 billion active parameters per token.
The Trinity Large model was particularly notable for its cost-efficiency, being built for approximately $20 million, a figure that challenged the industry's belief that training new models requires billions of dollars. The startup will use the funding to develop larger models, improve reasoning capabilities, and create more efficient architectures that can run on smaller hardware.
Additionally, Arcee will expand its work with the US Department of Energy and its national laboratories, as well as develop new products to help organizations customize, evaluate, deploy, and operate open models within Vista Equity Partners' portfolio companies.
Written by urgent.news from Ventureburn's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
