Mem0, a memory infrastructure platform for AI agents, has secured $24 million across Seed and Series A rounds led by Basis Set Ventures, aiming to position itself as the default memory layer for AI agents and large language models (LLMs). The round drew participation from Peak XV Partners, Kindred Ventures, GitHub Fund, and Y Combinator.
Alongside venture investors, a slate of prominent technology leaders also joined the cap table: Scott Belsky, Dharmesh Shah, and CEOs or founders from major infrastructure companies, Olivier Pomel (Datadog), Paul Copplestone (Supabase), James Hawkins (PostHog), Thomas Dohmke (ex-GitHub), and Lukas Biewald (Weights & Biases).
What Mem0 offers?
Mem0 provides production-ready agent memory infrastructure designed to drop into developer stacks “with a few lines of code.” According to the company, its memory layer extracts and categorises information from interactions, applies decay and confidence metrics, resolves conflicting facts, and retrieves only the most relevant memories in context.
Founded in 2023, Mem0 reports growing developer adoption, citing 41,000 GitHub stars and 14 million Python package downloads. Usage has accelerated this year: API calls rose from 35 million in Q1 2025 to 186 million in Q3 2025.
Mem0 is described as the exclusive AI memory provider for Amazon Web Services (AWS) for its new Agent SDK, giving the startup a direct line into developers building agentic applications on AWS.
The company’s pitch
“Every agentic application needs memory, just as every application needs a database,” said Taranjeet Singh, co-founder and CEO of Mem0. The fresh capital, he added, will be used to make LLM memory as accessible and reliable as databases or authentication, arguing that teams that integrate rich user context from day one will ship the next wave of breakthrough applications.
The Significance?
- Memory as a primitive: Mem0 is positioning memory as a first-class building block for AI, akin to databases for apps, by offering a managed, programmatic memory layer for agents and LLM workflows.
- Developer-first packaging: The emphasis on quick integration and governed retrieval of relevant context meets a practical developer need: keeping agents stateful and consistent across sessions without hand-rolled infrastructure.
- Platform distribution via AWS: The exclusive role within AWS’s Agent SDK concentrates visibility where agent builders already are, potentially accelerating adoption.
With new funding and an AWS tie-in, Mem0 is doubling down on the claim that “every agent needs memory.” The company’s near-term focus, based on its own description, is to translate rising developer interest into production usage at scale, underpinned by its relevance-aware retrieval, conflict resolution, and confidence/decay scoring.
Also Read: How Veranda Is Transforming Education Across India
