Engram, an AI ‘memory’ startup aiming to slash token spend, raises $98 million

Engram, an eight-month-old startup building “learned memory” systems to make enterprise AI cheaper, raised $98 million from General Catalyst, Kleiner Perkins and Sequoia, with OpenAI co-founder Andrej Karpathy participating. The 13-person company says its software recalls organization-specific workflows to cut token usage, claiming parity with frontier models while using up to 100 times fewer tokens. Early customers include Microsoft, Notion and legal AI firm Harvey. Engram plans to spend the funds on compute and hiring as companies seek to rein in rising AI inference costs. CEO Dan Biderman said the product favors specialized, context-rich accuracy over broad general capabilities.

Read more