mem0 Raises $24M to Revolutionize AI Memory Layer

The Rise of Mem0: A New Approach to AI Memory

Taranjeet Singh, a serial entrepreneur with six prior ventures, has introduced Mem0, potentially his most impactful creation to date. His previous companies experienced a range of outcomes, from failures to moderate successes.

Addressing a Core Limitation of Large Language Models

The fundamental challenge Mem0 addresses is the inability of current large language models (LLMs) to retain information from previous interactions. Unlike human conversations, which can be resumed after interruptions, AI models typically begin each exchange anew.

Mem0: A "Memory Passport" for AI

Singh describes Mem0 as a “memory passport,” enabling AI memory to accompany users across various applications and agents, functioning similarly to how email or login credentials operate today. The company, founded in January 2024, has successfully secured $24 million in funding.

Funding and Backers

This funding includes $3.9 million in previously undisclosed seed capital and a subsequent $20 million Series A investment. Basis Set Ventures spearheaded the Series A round.

Additional participation came from existing investors Kindred Ventures, who led the seed round, and Y Combinator. New investors include Peak XV Partners and the GitHub Fund.

Notable Angel Investors

A distinguished group of angel investors has also contributed to Mem0’s success. These include Dharmesh Shah (HubSpot), Scott Belsky (formerly CPO of Adobe), Olivier Pomel (Datadog), Thomas Dohmke (ex-CEO GitHub), Paul Copplestone (Supabase), James Hawkins (PostHog), Lukas Biewald (Weights & Biases), Brian Balfour (Reforge), Philip Rathle (Neo4j), and Jennifer Taylor (former president, Plaid).

Early Traction and Adoption

The backing of prominent figures in the software industry highlights the potential of Mem0. The four-person team has demonstrated significant early traction with developers.

The open-source API has garnered over 41,000 GitHub stars and exceeded 13 million Python package downloads. API calls processed by Mem0 rose from 35 million in Q1 2025 to 186 million in Q3, representing approximately 30% month-over-month growth.

Expanding Cloud Service and Partnerships

Over 80,000 developers have registered for Mem0’s cloud service. The platform now processes more memory operations than any other provider and serves as the exclusive memory solution for AWS’s new Agent SDK.

Singh’s Background and Journey

In early 2023, Singh was based in Bangalore, India. He began his career as a software engineer at Paytm, a leading Indian startup, and later became Khatabook’s initial growth engineer.

He departed Khatabook in late 2022, coinciding with the emergence of ChatGPT, and developed one of the first GPT app stores, attracting over a million users.

From Embedchain to Mem0

This experience led to the creation of Embedchain, an open-source project facilitating the indexing, retrieval, and synchronization of unstructured data. The project gained over 8,000 GitHub stars.

Singh proactively contacted numerous founders, investors, and engineers in Silicon Valley, ultimately securing meetings and a relocation to San Francisco.

Reuniting with a Co-Founder

In the U.S., Singh reconnected with Deshraj Yadav, his longtime friend and current co-founder and CTO. Yadav previously led the AI Platform at Tesla Autopilot. The pair had previously collaborated on EvalAI, an open-source alternative to Kaggle.

The Pivotal Insight

While developing Embedchain, they launched a meditation app inspired by Sadhguru. User feedback consistently highlighted the app’s inability to remember user progress, prompting a pivot from Embedchain to Mem0.

The Growing Importance of AI Memory

The concept of AI memory is gaining prominence, with companies like OpenAI testing long-term memory features in ChatGPT and CEO Sam Altman suggesting persistent memory will be crucial for future hardware devices. Other AI labs are also exploring similar systems.

Mem0’s Unique Value Proposition

Singh contends that larger AI labs lack the incentive to create portable or interoperable memory systems, viewing memory as a competitive advantage. He believes that as LLMs become more standardized, memory will become a key differentiator.

He emphasizes that while consumers may benefit from personalized experiences within platforms like ChatGPT, developers require an open and neutral solution like Mem0 to build applications with persistent user data.

A Vision for a Shared Memory Network

“We aim to empower developers to deliver personalized experiences from the outset through a shared memory network,” Singh stated. “Our goal is to become the ‘Plaid for memory,’ and we are currently focused on developing the most robust memory product available.”

Technical Capabilities of Mem0

Mem0’s framework enables developers to store, retrieve, and refine user memory across diverse models, applications, and platforms. It is compatible with various LLMs, including OpenAI, Anthropic, and open-source alternatives, and integrates seamlessly with frameworks like LangChain and LlamaIndex.

Applications and Customer Base

Developers are utilizing Mem0 to create applications that learn and adapt with each interaction, such as therapy bots that recall past conversations, productivity agents that remember user habits, and AI companions that evolve over time. Mem0’s customers include both individual developers and enterprise teams building copilots and automation tools.

Investor Confidence

“We invested in Mem0 early on – even before YC – because memory is fundamental to the future of AI,” said Lan Xuezhao, founder and partner at Basis Set Ventures. “We are increasing our investment as the team continues to address a challenging and vital infrastructure problem: enabling AI systems to build lasting, contextual memory.”

Competitive Landscape

Other early-stage startups in the AI memory space include Supermemory, Letta (backed by Felicis), and Memories.ai.