From Mexico — don’t lead with “another billion-dollar raise.” Arcee AI’s story is the cost thesis: an open-weight Trinity family trained lean, then a Series B that values the company at more than $1 billion. Distinct from today’s LawZero sovereign safety funding on X.
Trinity Large — architecture first
Per Arcee’s blog and GlobeNewswire: Trinity Large is a 400B-parameter sparse MoE with 13B active parameters per token — scaled in about six months from a 4.5B dense start. The company says its entire 2025 model lineup, including Trinity Large, was built for approximately $20 million (salaries, compute, data, infra). Fortune repeats the ~$20M train-four-models framing. Label that as company/Fortune — not my audit.
Series B investors — raise size = Fortune source
Arcee (and GlobeNewswire) put the Series B as led by Vista Equity Partners, Cambium Capital, and Emergence Capital, with participation from AI10 Ventures, Hitachi, IAG, M12, P7, and Wipro. The company values itself at more than $1 billion. Exact raise amount is not disclosed by Arcee; Fortune says a source familiar with the matter put it at at least $150 million, with a ~$1B pre-money frame — attribute that to Fortune, don’t invent a firm close.
Where the money goes — as company
Per Arcee: proceeds accelerate the next Trinity generation (already in training), expand work with the U.S. Department of Energy and national labs — naming Genesis-Science-1 as a starting point — and grow products to customize, evaluate, deploy, and operate open models. I’m not inventing DOE contract terms beyond that.
My takeaway
From Mexico, the narrow read: Arcee’s Series B is the open-weight efficiency pitch — Trinity Large at 400B MoE / 13B active after a ~$20M 2025 lineup cost, valued >$1B, with Fortune’s ≥$150M source line labeled — and proceeds pointed at next Trinity, DOE science, and open-model ops. Sources: Arcee, GlobeNewswire, Fortune.
