# The Startup That Wants to Solve AI Agents' Memory Problem

*Graphon is developing technology that connects documents, images and videos to allow artificial intelligence agents to remember information over time • Its customers already include HP and the San Francisco 49ers football team*

- Published: 2026-11-10T21:17:43.193Z (Tuesday, November 10, 2026)
- Updated: 2026-11-12T20:21:13.426Z
- By: TV10 Desk
- Section: Startups
- Canonical: https://tv10.news/news/tech/startups/127054/startup-solving-ai-agents-memory-problem/
- Hebrew original (TV10): https://tv10.co.il/news/tech/startups/127054/
- Publisher: TV10 Global (TV10, Israel’s economic channel). Translated from Hebrew with AI and an automated accuracy check.

Artificial intelligence agents can already write code, search databases and perform complex, multistep tasks. **But they still have a significant weakness:**They struggle to remember previous conversations, retain information over time and navigate large amounts of constantly changing data. Startup Graphon wants to solve this problem and has raised $35 million to do so.

The San Francisco-based company completed a Series A funding round led by venture capital firm RRE Ventures, with participation from Energy Impact Partners, Hitachi Ventures, Gaia VC and HP. Earlier this year, the company also completed a Seed funding round of about $8 million.

![Fundraising, illustration](https://tv10.news/media/2026/11/0a/0ac6f4f99efa83fe74b919baa7d2d1093f64f1c8b6ed8a2e44ce811626d8796e.jpg)

Fundraising, illustration | Photo: Shutterstock

Graphon is developing software that allows [AI agents](https://tv10.co.il/articles/סוכני-ai-ניסו-לפרוץ-לאתרי-ממשל-בארהב-ובק) to store information, retrieve it and understand the connections within it over time. Unlike conventional search systems, which treat documents and data as separate pieces of information, the company's technology maps the connections among documents, videos and images, including the source of the information and when it was updated.

According to the company's founder and CEO, Arbaz Khan, traditional AI search systems are like a book whose pages have been torn out, crumpled and thrown in the trash. When trying to find information in them, the relevant pieces must be located again, without necessarily preserving the context in which they were created.

Graphon instead seeks to create an interconnected information network that will allow agents to understand not only what was said, but also how the information relates to other details and when it changed.

The technology is intended for both independent developers and companies with large databases, mainly in the industrial, financial services and health care sectors. According to the company, the system can be connected to existing AI agents, including [Claude](https://tv10.co.il/articles/משתמשים-ב-claude-בעבודה-הבוס-שלכם-יוכל-לעקוב-אחריכם)and [GPT](https://tv10.co.il/articles/openai-משיקה-את-gpt-5-6-לאחר-שינויים-רבים-של-ממשל), and used to store and retrieve large amounts of information.

![ChatGPT](https://tv10.news/media/2026/11/74/74bd19205839c8e6a89db10666fe2eb885294f0429639683ca77d19a5d1f4435.jpg)

ChatGPT | Photo: Shutterstock

Customers already using Graphon's technology include HP and the San Francisco 49ers American football team. The company currently employs about 20 people and plans to use the funding mainly to expand its engineering and research teams, alongside hiring employees in sales and business development.
