The archive · AI & Models · Strategic decision · 2023–2025
Mem0 bets every AI agent needs a portable memory layer
Mem0 raised $24M to become the default memory layer for AI agents: 41k GitHub stars, 14M downloads, and AWS picked it for its Agent SDK.
Mem0
What the business is
Open-source memory layer for AI agents: an API and SDK that stores, retrieves and evolves user memory across models, apps and agents, plus a managed cloud platform.
Starting capital：$24M total: $3.9M previously unannounced seed plus $20M Series A led by Basis Set Ventures, with Peak XV Partners, Kindred Ventures, GitHub Fund and Y Combinator (announced Oct 28, 2025)
How it started
Taranjeet Singh, an ex-Paytm and Khatabook engineer who built Embedchain, teamed up with Deshraj Yadav, who had led the AI platform at Tesla Autopilot, and launched Mem0 in January 2024. While experimenting with Embedchain they built a meditation app that went viral in India, and the recurring user complaint — 'the app doesn't remember me' — became the product.
What happened
Mem0 grew through open source and Y Combinator: the framework integrated with LangChain, LlamaIndex, CrewAI, Flowise and Langflow, and AWS selected Mem0 as the memory provider for its Agent SDK. By the funding announcement in October 2025 the open-source project had surpassed 41,000 GitHub stars and 14 million Python downloads, API volume jumped from 35 million calls in Q1 2025 to 186 million in Q3 2025, and more than 80,000 developers had signed up for the cloud service. The $24M round — $3.9M seed plus $20M Series A led by Basis Set Ventures — brought in Peak XV Partners, GitHub Fund and angels like Dharmesh Shah, Scott Belsky and Olivier Pomel.
How it ended up
Scaling: with $24M, Mem0 plans to expand its engineering team, harden memory for enterprise use cases and build partnerships with AI platforms; the GitHub repo remained actively developed through 2026, shipping a new memory algorithm in April 2026.
Background
Mem0 is a memory infrastructure platform for AI agents, launched in January 2024 by Taranjeet Singh and Deshraj Yadav. Singh, an ex-Paytm and Khatabook engineer, had built Embedchain, an open-source project for indexing and retrieving unstructured data; Yadav had led the AI platform at Tesla Autopilot. The idea came from a meditation app the pair built while experimenting with Embedchain: it went viral in India, and users kept saying the same thing — the app didn't remember them.
Mem0's bet is that LLM memory should be a portable, model-agnostic layer developers own, not a feature locked inside any single AI lab's product. Its framework lets developers store, retrieve and evolve user memory across models, applications and platforms, working with OpenAI, Anthropic or any open-source model and integrating directly with LangChain and LlamaIndex. Singh describes the ambition as 'Plaid for memory' — a memory passport that travels with you across apps and agents.
The bet validated fast. By the October 2025 funding announcement, the open-source project had surpassed 41,000 GitHub stars and 14 million Python downloads; API calls grew from 35 million in Q1 2025 to 186 million in Q3 2025; more than 80,000 developers had signed up for the cloud service; and AWS selected Mem0 as the memory provider for its Agent SDK. The $24M round — $3.9M previously unannounced seed plus a $20M Series A led by Basis Set Ventures, with Peak XV, Kindred Ventures, GitHub Fund and Y Combinator — is funding a bigger engineering team and deeper enterprise memory features.
What has to be true
- LLMs forget everything between sessions, and big AI labs have little incentive to make memory portable; Mem0 offered a neutral, model-agnostic layer instead.
- Open-source adoption (41,000+ stars, 14M downloads) and 186M quarterly API calls showed real developer demand, not just hype.
- AWS picking Mem0 as the memory provider for its Agent SDK was a structural validation of the infrastructure bet.
- The pivot came from observed user pain — a viral app that users wanted to remember them — rather than a top-down market thesis.
What can be applied
Pivot where users complain: feedback from a viral side project — 'it doesn't remember me' — exposed a platform gap bigger than the app, and a memory layer built for that pain became a $24M business.
Aftermath
As of the October 2025 funding announcement, Mem0 is scaling with a four-person team: 41,000+ GitHub stars, 14M+ Python downloads, 186M API calls in Q3 2025, more than 80,000 cloud signups, and exclusivity as the memory provider for AWS's Agent SDK. The founders plan to grow the engineering team, deepen enterprise memory features and extend platform partnerships; the GitHub repository remained actively developed through 2026, shipping a new memory algorithm and benchmarks in April 2026.
Sources
- Mem0 raises $24M from YC, Peak XV and Basis Set to build the memory layer for AI apps
- Mem0 Raises USD 24 Mn to Strengthen AI Memory Infrastructure
spotted an error? The archive wants to know.
Your turn
You just read one. Describe what you are building, and see who is betting on the same thing.
Free account · 3 free questions · no card