Arcee AI has completed a $150 million Series B at a $1 billion pre-money valuation to expand its open-weight artificial intelligence models and develop a system exceeding one trillion parameters. The San Francisco-based company was founded in 2023 by Mark McQuade, Jacob Solawetz and Brian Benedict. It initially focused on fine-tuning and optimizing existing models before shifting to building models from scratch. Arcee’s Trinity family includes models ranging from 6 billion to 400 billion total parameters, with the flagship model activating about 13 billion through a mixture-of-experts architecture. The company says it trained four models for about $20 million using 2,048 Nvidia Blackwell B300 GPUs and a team of approximately 30 people. The round was led by Vista Equity Partners, Cambium Capital and Emergence Capital, with participation from Microsoft’s M12, AI10 Ventures, Hitachi, IAG, P7 and Wipro. Earlier disclosures also listed Hitachi Ventures, Samsung Next and Prosperity7 Ventures. Arcee’s Apache 2.0-licensed models target enterprises in regulated sectors that need to run and customize AI on their own infrastructure. The new capital will support additional models and products, Arcee’s partnership with the U.S. Department of Energy, work with Vista’s portfolio companies and the planned trillion-parameter system.