2 hours ago
20
Redis, the company synonymous with the caching layer that prevented web applications from buckling under traffic, is making a sharp pivot deeper into AI infrastructure. On Monday, it launched Iris, a context and memory platform purpose-built for AI agents, targeting what it sees as a fundamental mismatch between how agents consume data and how most retrieval systems were designed to serve it.
The core thesis is straightforward: AI agents make orders of magnitude more data requests than human users, but most retrieval pipelines were built for the human-scale problem. Iris is Redis’ attempt to close that gap before it becomes the bottleneck that stalls enterprise AI adoption.
What Iris actually does
LLMs are inherently stateless. Every interaction starts from scratch unless something external provides continuity. That external something is what Iris is designed to be.
The platform sits between an AI agent and the data it needs to act. It consolidates three capabilities that enterprises have typically had to stitch together from separate tools: a Context Retriever, Agent Memory, and Data Integration.
The Context Retriever handles real-time data fetching, pulling in both structured and unstructured information so an agent can ground its responses in current facts rather than whatever its training data happened to include. Agent Memory provides both short-term and long-term persistence, meaning an agent can recall what happened in prior sessions, track evolving user preferences, or maintain state across multi-step workflows. The Data Integration layer, which Redis calls RDI, acts as a real-time data loader that keeps the underlying information fresh.
Why this matters for AI infrastructure
Iris also arrives alongside a new Flex SSD-based version of Redis, which suggests the company is thinking about cost efficiency alongside performance. Running everything in-memory is fast but expensive. An SSD tier could make it feasible for enterprises to maintain larger context windows and longer agent memories without blowing up their infrastructure budgets.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.
English (US) ·