Anysphere raises $2.3B Series D

Background

Anysphere is the company behind Cursor, an AI coding tool that puts model-generated code directly into the place where developers write and edit software. The Series D announced on 13 November 2025 came nearly two years after the company's seed round, and the company's own announcement framed the new money as funding for research toward Cursor's next products^[1].

What happened

Anysphere announced a $2.3 billion Series D on 13 November 2025^[1]. The round valued the company at $29.3 billion post-money^[1]. Existing investors Accel, Thrive and Andreessen Horowitz took part, and were joined by Coatue, NVIDIA and Google^[1].

Key details

Alongside the financing, the company gave a set of figures about its own scale. It said it had grown to more than 300 engineers, researchers, designers and operators^[1]. It said it had crossed $1 billion in annualized revenue and counted millions of developers as customers^[1]. It also said its in-house models generate more code than almost any other large language models in the world^[1].

Taken together, the three figures describe three different things: how many people the company employs, how much money it brings in, and how much it relies on models it built itself. Annualized revenue is a measure that takes revenue over some period and scales it to a full year; crossing $1 billion on that basis means the business is collecting money from customers at a rate that would reach ten figures over twelve months. A staff of more than 300 is small next to that revenue figure, which is one reason AI coding tools have drawn so much attention: a comparatively compact team can serve a very large user base.

Why it matters

The numbers describe a company that has moved from a small startup to one with a large headcount, a large user base and substantial revenue in a short period. The investor list is notable in two directions: it keeps the venture firms that backed earlier rounds, and it adds a chipmaker and a major cloud and AI company, both of which sit close to the infrastructure that AI coding tools depend on^[1].

The stated purpose of the money — research toward the next generation of products — indicates that the company intends to keep building its own models rather than only wrapping those of others^[1]. That claim is backed by the company's assertion about how much code its in-house models generate^[1]. For developers, the practical stake is what the next round of products does inside the editor; for the wider field, the round is a signal about how much capital is flowing toward tools that write code alongside people rather than replacing the editor outright.