Ryan Trimberger, 4Minds
In this interview from theCUBE + NYSE Wired: Mixture of Experts, Ryan Trimberger, co-founder and chief executive officer of 4Minds.ai, joins theCUBE's John Furrier to discuss how GPU-accelerated knowledge graphs are solving enterprise AI's data fragmentation problem. Trimberger explains that enterprises moved quickly past search and retrieval once compliance and scale exposed the limits of RAG across fragmented datasets, from PDFs and spreadsheets to Salesforce and Oracle. He details how 4Minds' knowledge graph uncovers complex relationships across these disparate sources while a proprietary pruning method prevents the bloating that typically slows large graphs, preserving RAG-level speed with far greater intelligence. The conversation also explores 4Minds' rapid growth, including a $500 million Series A valuation — the fourth-largest in Texas history — now fueling a go-to-market team sized for a Series D company, with partnerships spanning Microsoft, Amazon, Google, Cisco, HPE, Deloitte and Accenture. Trimberger notes that over 90% of the world's data is private, representing the true alpha enterprises must protect rather than hand over to public models. He explains how knowledge graphs cut hallucinations, improve token efficiency and reduce latency, especially in regulated industries like financial services and healthcare where speed is itself a competitive moat. From deploying open-weight models on customers' own infrastructure to scaling specialized AI research teams, Trimberger outlines why enterprises are just beginning to unlock the intelligence locked inside their own data.