AI memory startup focused on cutting token costs raises $98 million (www.cnbc.com)

AI memory startup Engram has successfully raised $98 million to enhance model efficiency while drastically reducing operational costs. Founded just eight months ago, Engram claims its innovative models can outperform or equal those of leading labs while using up to 100 times fewer tokens—the essential currency for AI queries. This significant breakthrough comes at a time when organizations are increasingly looking to manage the rising expenses associated with advanced AI technologies. Notable backers of the funding round include General Catalyst, Kleiner Perkins, and OpenAI co-founder Andrej Karpathy. Engram’s technology enables AI models to recall specific organizational workflows and contexts, allowing for smarter, more relevant responses without the overwhelming data input typically required. As AI systems grow more sophisticated and expensive, Engram aims to flip the narrative by providing cheaper alternatives that maintain performance. Co-founder and CEO Dan Biderman emphasizes the need for AI to mimic human-like memory capabilities while also recognizing that their models prioritize specialization to deliver enhanced intuition over generalized intelligence. Engram’s approach not only signals a strategic pivot in AI cost management but also opens new avenues for companies seeking tailored AI solutions.