Graphon Labs, a San Francisco startup that builds memory for AI agents, has raised a $35M Series A led by RRE Ventures.
HP, Energy Impact Partners, Hitachi Ventures and Gaia VC also invested, Graphon said. The round brings total funding to $43.3M. The company left stealth less than six months ago with $8.3M, so the new round is roughly four times that opening amount.
The problem it targets is familiar to anyone who has used a chatbot. AI agents, meaning software that takes actions on a user's behalf, mostly work from what fits inside a single conversation. Unless developers add storage on top, an agent tends to start each session with a blank slate. Graphon's business is selling that missing layer.
The size of the round is the interesting part. Investors are betting that memory can stand as its own product rather than a feature each model provider builds in. That is not a given. Capabilities that start as third-party add-ons have a habit of being absorbed by the platforms they sit on.
The investor list adds a second thread. HP and Hitachi Ventures are tied to large hardware and industrial companies, which sometimes means customers or distribution. Graphon has not disclosed pricing, customers, its valuation or what those investors plan to do with the company, so that reading is an inference, not a fact.
The larger question is whether agent memory ends up as shared infrastructure that many agents plug into, or as a quiet checkbox inside whichever model a company already pays for.