Engram Launches With $98 Million, Testing AI Memory Layer in Microsoft 365

3 articles · Updated · PR Newswire · Jun 23

$98 million-backed Engram emerged from stealth on Tuesday, pitching a learned “organizational memory” layer that lets enterprise AI retain company-specific knowledge instead of relearning it on every query.
Engram says that approach can match or beat frontier models while using up to 100x fewer tokens by studying documents in advance and compressing them into reusable memory.
Microsoft is Engram’s anchor early partner, testing the models inside Microsoft 365 and providing GPU capacity across Dapple and Azure to support training at scale.
Notion and Harvey are also integrating the system into enterprise agents, with Notion saying early tests show near-frontier quality at roughly an order of magnitude lower token use.
Founded by researchers from Stanford, Berkeley and Cornell, Engram is targeting a fast-growing enterprise AI market where token costs, long-running agents and control over proprietary knowledge are becoming central concerns.

Sources

Left50%

Center50%

PR Newswire3h ago

Engram Launches with $98M Funding for AI Organizational Memory Solution

CNBC4h ago

Engram Raises $98 Million to Cut AI Token Costs

Crypto Briefing4h ago

AI Memory Startup Engram Raises $98M to Boost Model Efficiency, Cut Token Costs

4 Sources

Can Engram's memory tech survive as big AI labs race to solve the same cost problem?

With AI data centers consuming more power, can software alone solve the industry's massive energy problem?

From Token Cost Crisis to Memory-Centric AI: Engram’s 1 Million Token Context Window Transforms Enterprise AI

Overview

The AI industry is facing a major cost crisis as companies rapidly adopt AI tools and autonomous agents, causing token consumption and overall expenditures to soar. Even though per-token prices are falling, businesses are struggling to control budgets and understand their true return on investment. High-profile cases like Uber exhausting its AI budget early and Microsoft revoking developer licenses highlight the financial strain. Analysts predict that token costs could soon make up a significant portion of company expenses, marking a structural transformation in how organizations must plan and justify AI spending. This urgent challenge is driving demand for innovative, cost-efficient solutions.

...