--- title: "In January this year, the \"AI programming star\" Cursor, which is \"full of glory,\" discovered that the market may be gone" type: "News" locale: "en" url: "https://longbridge.com/en/news/278204585.md" description: "The AI programming tool unicorn Cursor held a \"wartime mobilization\" all-hands meeting in January this year. Due to the leap in AI model programming capabilities, developers can directly issue commands to the intelligent agents, fundamentally shaking the core logic of its \"human-machine collaborative editor.\" Cursor is breaking through along two paths: developing its own proprietary programming model to reduce costs and accelerating its transformation towards the enterprise market, betting on the future of multi-agent collaboration platforms" datetime: "2026-03-07T08:03:39.000Z" locales: - [zh-CN](https://longbridge.com/zh-CN/news/278204585.md) - [en](https://longbridge.com/en/news/278204585.md) - [zh-HK](https://longbridge.com/zh-HK/news/278204585.md) --- > Supported Languages: [简体中文](https://longbridge.com/zh-CN/news/278204585.md) | [繁體中文](https://longbridge.com/zh-HK/news/278204585.md) # In January this year, the "AI programming star" Cursor, which is "full of glory," discovered that the market may be gone The most dazzling unicorn in the AI programming tool sector, Cursor, is facing a strategic transformation that could determine its survival. On January 5th of this year, Cursor employees returned to the company from their holiday to find themselves in an all-hands meeting themed "wartime status." The trigger for this emergency meeting was a disturbing technological reality: **the programming capabilities of AI models have become so powerful that human oversight of code line by line is no longer necessary; developers can directly issue high-level instructions to autonomous agents and receive complete functionalities, fundamentally shaking the core product logic of Cursor's "human-machine collaborative code editor."** The company immediately set its **highest priority task as "building the best programming model"—not the best packaging tool, but the model itself.** This shift occurred at what seemed to be Cursor's unstoppable peak. The company had an annualized revenue of approximately $100 million at the beginning of 2025, which had surpassed $1 billion by November, with the latest round of financing valuing it at nearly $30 billion. **The sudden change in market dynamics is forcing Cursor to rewrite its business logic:** shifting from consumer subscriptions to enterprise contracts, moving from reliance on external models to developing proprietary programming models, and transitioning from a code editor to a multi-agent collaboration platform. Whether this transformation can succeed will determine if this company, founded just three years ago, can maintain its leading position in the AI programming arena. ## A "wartime mobilization" triggered by the holiday The all-hands meeting on January 5th was prompted by employees' firsthand experience with Anthropic's latest model, Opus 4.5, during the holiday. **The programming capabilities of this model have evolved to a new stage:** developers no longer need to collaborate line by line with AI assistants in the editor, but can **issue high-level instructions to autonomous agents and directly receive complete functional modules, sometimes even finished products.** **This poses a direct threat to Cursor.** CEO Michael Truell had previously defined Cursor in 2024 as "the Google Docs for programmers"—a collaborative editor where humans and AI refine code together. But **if AI no longer requires human collaborators, the very existence of the editor becomes precarious.** Cursor's management warned in the all-hands meeting that **the coming months would be turbulent, projects might be halted, and priorities would be rearranged.** The company's highest priority task was named "P0 #1," with only one goal: to build the best programming model. Internally, this moment was referred to as a "cleansing." ## The rapid myth from zero to ten billion Cursor was founded in 2022 by four MIT alumni, initially focusing on helping mechanical engineers design components, but quickly pivoted due to a lack of domain expertise, ultimately launching a code editor product that ignited the market. The speed of this company's rise has drawn industry attention. A year ago, Cursor pushed its annualized revenue to $100 million with just 20 employees and no sales team, and the viral spread of its product attracted the attention of top venture capital firms such as Accel, Andreessen Horowitz, and Thrive Capital In 2025, Anthropic listed Cursor as its largest client and granted it priority access to its models, forming a "co-opetition" relationship. By November 2025, Cursor's annualized revenue surpassed $1 billion, with its latest financing valuation approaching $30 billion, ranking among the top 20 most valuable private companies globally. The company's workforce expanded to about 400 people, occupying four buildings in the North Beach neighborhood of San Francisco, even transforming the bus stop advertising space between office buildings into a wall displaying employee names. **Despite facing skepticism regarding its technological direction, Cursor's financial data continues to show explosive growth.** According to Forbes, citing informed sources, Cursor's annualized revenue doubled within three months, now exceeding $2 billion. Data from corporate credit card companies Ramp and Brex also indicate that its revenue continued to grow in February, although Ramp noted a slight decline in adoption rates for corporate AI product procurement. However, the expansion speed of competitors is equally astonishing. Anthropic's Claude Code surpassed $1 billion in annualized revenue within six months, reaching $2.5 billion last month, outpacing Cursor. OpenAI CEO Sam Altman stated that its programming agent Codex exceeded 1 million downloads in the first week after its relaunch in April 2025. In the rapidly changing AI field, market momentum can shift at any moment. ## Competitors' "Dimensionality Reduction Attack" Cursor's crisis stems from a fundamental shift in the AI programming paradigm. Anthropic had warned Cursor about the development plan for Claude Code as early as last year—this is a command-line tool with a minimalist interface that allows developers to quickly deploy a large number of programming agents. At that time, this product seemed not to directly compete with Cursor's code editor. But the situation has changed dramatically. Several startup founders have indicated that this shift is profound. Andrew Hsu, co-founder and CTO of the AI language learning app Speak, stated, "**This is the most significant and fundamental transformation in the history of software development.**" All 50 engineers at his company have shifted to using programming agents (primarily Claude Code, with some scenarios using Codex), and features that previously took months to complete can now be delivered in weeks, reducing Cursor's role in their workflow. In February of this year, the mortgage service startup Valon saw over 90 employees cancel their Cursor subscriptions in favor of using Claude Code's agents for end-to-end automation tasks, including data migration and bug fixes. CEO Andrew Wang stated that the speed of completing related tasks has increased by "10 times." Jerry Murdock, co-founder of Insight Partners, candidly remarked on the 20VC podcast, "Most of the companies I mentioned... their view is that Cursor is already outdated today." ## Self-Developed Models and Enterprise Market: Two Paths to Breakthrough In the face of pressure, Cursor is seeking breakthroughs along two paths. **In terms of self-developed exclusive programming models**, Cursor has about 20 AI researchers focused on developing the company's proprietary Composer series models. These models are based on Chinese open-source models such as DeepSeek, Kimi, and Qwen, optimized through additional training and reinforcement learning based on Cursor's proprietary data. **The core logic is:** a specially trained small model can effectively compete with large cutting-edge models in terms of cost and performance. Currently, Composer 1.5 has become the second most popular model on the platform, with operating costs significantly lower than those of Anthropic's large models. **However, cost pressure remains prominent.** According to media reports, an insider revealed that internal estimates at Cursor show that Anthropic's subsidies for Claude Code are extremely aggressive— a subscription plan costing $200 per month can actually consume about $5,000 worth of computing power. Cursor also subsidizes some users, with consumer-level subscriptions operating at a negative gross margin, but enterprise-level plans have achieved a positive gross margin. **In the enterprise market**, Cursor is accelerating its focus on large clients. Although the signing cycle for enterprise contracts is longer, the customer churn rate is extremely low— according to an insider, Cursor has only lost one or two enterprise clients to date. Currently, about 60% of the company's revenue comes from enterprise clients, and the sales team has signed contracts with large clients such as Meta and Nvidia. The internal structure of the company has also adjusted accordingly, with half of the employees now focused on market expansion functions. In contrast, enterprise contracts accounted for only 13.6% of Cursor's annualized revenue last November, and the rapid increase in this proportion reflects a clear shift in the company's strategic focus. ## "Delete Product": Betting on a Multi-Agent Future One straightforward directive in Cursor's internal values is: "Delete Product"— this is the company's clear recognition of its future: code editors will eventually give way to programming agents. Last week, Cursor announced a major update to its "Cloud Agents" product, allowing multiple agents to handle different tasks simultaneously in their own independent workspaces and record their work processes. The company is exploring a mode referred to as "grind mode"— allowing a single developer to coordinate hundreds of agents working together. This vision faces complex technical challenges: **how to assign exclusive roles to each agent, and how to prevent agents from experiencing efficiency declines when they "perceive many peers"**— this phenomenon is quite similar to "free-riding" behavior in human teams. **Cursor's management bets on a differentiated advantage:** enterprise clients will increasingly value products that do not rely on a single model supplier, especially in the current landscape where the capabilities of various models are constantly changing. Whether this judgment holds true will largely determine whether Cursor can maintain its position in this AI programming war ### Related Stocks - [Global X Cloud Computing ETF (CLOU.US)](https://longbridge.com/en/quote/CLOU.US.md) - [Microsoft Corporation (MSFT.US)](https://longbridge.com/en/quote/MSFT.US.md) - [OpenAI (OpenAI.NA)](https://longbridge.com/en/quote/OpenAI.NA.md) - [State StreetSPDRS&PSftwr&SvcsETF (XSW.US)](https://longbridge.com/en/quote/XSW.US.md) - [iShares Global Tech ETF (IXN.US)](https://longbridge.com/en/quote/IXN.US.md) - [iShares Expanded Tech-Software Sect ETF (IGV.US)](https://longbridge.com/en/quote/IGV.US.md) ## Related News & Research - [Timekettle Makes Its First Appearance at MWC 2026, Highlighting the Highly Responsive W4 AI Interpreter Earbuds](https://longbridge.com/en/news/277373612.md) - [Malware-laced OpenClaw installers get Bing AI search boost](https://longbridge.com/en/news/277830709.md) - [The 'dean of valuation' pours cold water on the Citrini AI doomsday scenario](https://longbridge.com/en/news/277924974.md) - [CoreWeave's new AI investment: Chance the Rapper](https://longbridge.com/en/news/277758078.md) - [Zetrix AI Completes Private Placement of Shares](https://longbridge.com/en/news/277724680.md)