Engram has emerged from stealth with $98 million in funding to build a learned memory layer for AI, backed by General Catalyst, Kleiner Perkins, Sequoia Capital, Factory, Modern, Amplify Partners, Neo, and notable angels and advisors including Assaf Rappaport, Andrej Karpathy, and Pieter Abbeel. The company aims to help enterprise AI systems understand each organization’s unique context, addressing the issue of AI systems repeatedly rereading documents and rediscovering institutional knowledge. Engram’s technology trains models to study an organization’s world and anticipate questions in advance, creating a compact and continuously improving memory unique to each customer. The company said its models can match or outperform frontier models while using up to 100 times fewer tokens, reducing cost and complexity for enterprises. Founded by AI researchers from Stanford, Berkeley, and Cornell, the team includes CEO Dan Biderman, CTO Sabri Eyuboglu, Jessy Lin, Jack Morris, Scott Linderman, and Chris Ré. Engram has early commercial traction, including a partnership with Microsoft to test models within Microsoft 365, with a commitment to GPU capacity across Dapple and Azure. Engram is also partnering with Notion and Harvey to bring its memory layer into their platforms.