LOGO

Cursor CEO on OpenAI & Anthropic Competition - Startup Strategy

December 9, 2025
Cursor CEO on OpenAI & Anthropic Competition - Startup Strategy

Anysphere's Focus on Feature Development, Not an IPO

Michael Truell, co-founder and CEO of Anysphere – the creator of the popular AI coding assistant Cursor – indicated Monday at Fortune’s AI Brainstorm conference that an initial public offering is not currently under consideration for the company.

Recent Growth and Investment

Having achieved an annualized revenue of $1 billion in November, and subsequently securing $2.3 billion in funding at a $29.3 billion valuation last month, Anysphere’s immediate priority lies in expanding its feature set.

Development of In-House LLMs

Cursor utilizes internally developed Large Language Models (LLMs) specifically designed to enhance its product capabilities. The company affirmed the existence of these models in a November blog post, stating they now generate more code than many other LLMs available.

Competing with Major LLM Providers

The discussion surrounding these models arose during Fortune’s event when Truell was questioned about his strategy for competing with leading LLM developers – such as OpenAI and Anthropic – who are also introducing AI coding solutions.

A Production-Ready Solution

Truell characterized competing coding products as “a concept car,” contrasting them with Cursor’s fully developed, production-ready automobile. He explained, “We integrate the best available intelligence from various providers, alongside our own product-specific models, to deliver a comprehensive and optimized experience.”

Past Acquisition Attempts and Pricing Adjustments

Anysphere’s reliance on external LLMs and its subsequent development of in-house models has been a topic of discussion within Silicon Valley venture capital circles. Earlier this year, OpenAI reportedly explored acquiring Anysphere, but the offer was declined.

Investors had expressed concerns about the financial viability of AI coding editors due to the substantial costs associated with LLM usage. In response, Cursor transitioned to a usage-based pricing model in July, directly passing API fees onto its users. This change led to some dissatisfaction among customers due to unexpectedly high bills.

Addressing Pricing Concerns

Addressing the pricing adjustments, Truell stated, “As Cursor has evolved from a tool for quick code snippets to a platform for extensive work, the pricing model needed to adapt, mirroring a shift towards consumption-based billing across the industry.”

Focus on Enterprise Cost Management

The company is currently developing cloud-computing-like cost-management tools to assist enterprises in monitoring usage and controlling expenses related to their engineers’ AI tool utilization.

Key Development Areas for the Next Year

Anysphere is concentrating its efforts on two primary areas for the coming year. These include enhancing complex agentic functions and prioritizing team-based workflows.

Advanced Agentic Capabilities

The goal is to enable Cursor to autonomously handle end-to-end tasks, even those that are difficult to specify but require significant effort. A prime example is automated bug fixing.

Truell envisions Cursor resolving complex bugs that might take weeks of manual effort and countless code executions, completing the entire process independently.

Serving Teams as a Core Unit

The company is shifting its focus to serving teams as the fundamental unit of operation, indicating a growing emphasis on its enterprise business.

Expanding Beyond Code Generation

Beyond code generation, Anysphere intends to expand Cursor’s capabilities to encompass more aspects of the software development lifecycle. The existing code review product, used by some customers to analyze all pull requests – whether generated by AI or humans – serves as a model for this expansion.

Industry Competition and Open Standards

Major competitors are also actively developing advanced agentic capabilities. Amazon recently launched a coding tool capable of running for extended periods. Furthermore, a consortium of AI leaders, including Anthropic, OpenAI, Microsoft, and AWS, has formed under the Linux Foundation to establish open-source standards for agentic interoperability.

Maintaining a Competitive Position

While these developments may not immediately position Anysphere ahead of its primary competitors, they are expected to ensure the company remains a significant player in the AI coding landscape.

#Cursor#OpenAI#Anthropic#AI startups#AI competition#CEO interview