I'm not OP but that seems like a pretty reasonable assumption. LLM dominance is basically a US thing (Europe is handicapped by the EU and China is handicapped by hardware), and there are only a few companies that are actually competitive in LLM's (OpenAI, Gemini, xAI, Meta, Anthropic). I think top 5 is a safer bet but top four isn't an unreasonable thing to say confidently
xAI has the largest GPU cluster for AI training in the world, and they regularly produce top-ranking models. Companies like mistral produce clever models but they’re functionally just not on the level of things like Grok