Mem0 Raises $24M Series A to Build Memory Layer for AI Agents

Mem0, a memory infrastructure platform for AI agents, today announced $24 million in combined Seed and Series A funding led by Basis Set Ventures, with participation from Peak XV Partners, Kindred Ventures, GitHub Fund, and Y Combinator. The round includes strategic investments from technology leaders Scott Belsky and Dharmesh Shah, alongside CEOs of major infrastructure companies including Olivier Pomel (Datadog), Paul Copplestone (Supabase), James Hawkins (PostHog), Thomas Dohmke (ex-GitHub), and Lukas Biewald (Weights & Biases).

Despite rapid advances in AI reasoning and decision-making, today’s AI agents face a fundamental limitation: they can’t remember. Users repeatedly paste the same context into ChatGPT, re-explain preferences to support bots, and watch coding assistants suggest the same rejected patterns. This lack of AI agent memory prevents personalization, continuity, and long-term learning across interactions. For developers, building AI memory from scratch consumes time and energy that could be spent on creating unique, differentiated applications. Many rely on workarounds like RAG pipelines or overloading context windows, which fail to provide true persistence and efficiency.

Mem0 provides production-ready agent memory infrastructure that developers can integrate with three lines of code, transforming stateless agents into systems with powerful personalization. Their memory layer intelligently extracts and categorizes information from interactions, applying metrics like decay and confidence, resolving conflicting facts, and retrieving only the most relevant memories in context. Since launching, the company has reached 41,000 GitHub stars and 14 million Python package downloads, with API calls growing from 35 million in Q1 to 186 million in Q3 2025

Mem0 was founded in 2023 by Taranjeet Singh and Deshraj Yadav who have deep expertise in AI infrastructure and developer tools. Taranjeet previously built Cookup AI, the first GPT App store — which scaled to over one million users, and also created a personalized AI tutor app which got featured at Google I/O conference. Deshraj led AI infrastructure at Tesla Autopilot, where he worked with Andrej Karpathy. Together, they have also built Embedchain, a popular RAG framework with over 10 million downloads, and EvalAI, an open-source alternative to Kaggle with more than 1,800 GitHub stars.

Taranjeet Singh, Co-founder and CEO of Mem0 said “Every agentic application needs memory, just as every application needs a database. We’re using this funding to become the default memory layer for AI agents and LLMs — making LLM memory as accessible and reliable as databases or authentication. The developers who integrate rich user context from day one will build the next generation of breakthrough applications.”

Lan Xuezhao, Founding Partner at Basis Set Ventures said “We backed Mem0 since the earliest days even before YC because memory is foundational to the future of AI. We’re thrilled to double down today as the team continues solving one of the hardest and most important infrastructure challenges: enabling AI systems to build lasting, contextual memory.”

Thousands of teams, from startups to Fortune 500 companies, now use Mem0 in production. Leading frameworks like CrewAI, Flowise, and Langflow have integrated it natively, and AWS has chosen Mem0 as the exclusive AI memory provider for its new Agent SDK.

“Memory infrastructure is becoming mission-critical for AI personalization. Imagine a world where all your memories, context and preferences are stored in a database — portable across apps through a single “memory sign-on”. We are excited to partner with Taranjeet and Deshraj at Mem0, which has seen adoption by thousands of developers in leading AI companies. Mem0 has the potential to be a n of 1 company.” Arnav Sahu, Partner, Peak XV

With this funding, Mem0 will expand its engineering team, enhance memory capabilities for complex enterprise use cases, and establish partnerships with major AI platforms and frameworks.

Developers can get started today at www.mem0.ai or explore the open-source Mem0 GitHub repository.