Claude and Clawdbot ignite the AI agent frenzy! "OpenAI rival" Anthropic revenue expected to increase 20% to $18 billion.

Media reports cited information from informed sources revealing that the leader in the generative AI field, Anthropic, known as the "OpenAI rival," predicted at an internal meeting in December 2025 that it would achieve a record-high revenue of $18 billion in 2026, representing a 20% increase from its previous summer forecast. It is also expected to reach around $55 billion next year - indicating that Anthropic believes its revenue will more than triple under the influence of its core AI application products such as Claude Code and Sonnet which have become popular worldwide. According to media reports, the AI application leader Anthropic, backed by heavy investments from one of the largest tech giants in the US Amazon.com, Inc. and Alphabet Inc. Class C, predicted a total revenue peak of $148 billion by 2029 in its most optimistic forecast, surpassing OpenAI's estimate by around $3 billion for that year. Anthropic did not immediately respond to media requests for comments. Sources indicated that a significant portion of the rapidly growing revenue will come from Anthropic's successful selling of access and permissions to its flagship AI applications to large enterprise customers through Application Programming Interfaces (API). The company has gained popularity in AI programming with its agent-based AI programming application Claude Code, which generated over $1 billion in annual revenue in November 2025. AI programming is currently one of the hottest technology sectors in the global AI application wave, transitioning from "personal developer tools" to "enterprise standard configuration." AI programming has gained popularity globally, signaling the arrival of a super wave of AI applications that cover various industries, leading to a significant productivity transformation comparable to the industrial revolution. Similar to the cash-burning OpenAI, Anthropic's costs are also accumulating rapidly. Anthropic revised its most optimistic forecast, delaying the time for achieving positive cash flow to 2028, a year later than previously expected. Anthropic estimates it will generate $2.2 billion in cash flow in 2028. Reports indicate that the higher cost of training and efficiently operating its large AI models contributed to this delay. In 2026, Anthropic expects to spend $12 billion on training models and around $7 billion on efficiently running AI large model technology for its paid users. In a financing round led by Singapore's Government Investment Corporation (GIC) and Coatue Management, this AI unicorn is raising over $10 billion at a pre-financing valuation of $350 billion. Last week, reports indicated that Anthropic's latest revenue run rate at the end of 2025 had already exceeded $9 billion. In November 2025, NVIDIA Corporation and Microsoft Corporation had pledged investments of up to $10 billion and $5 billion, respectively, in Anthropic. Microsoft Corporation has invested over $13 billion in OpenAI, giving it close to a 27% stake in this ChatGPT developer. Earlier this month, reports stated that Microsoft Corporation had become one of Anthropic's top clients and was expected to purchase Anthropic's proprietary AI large model technology at a rate of approximately $500 million per year to support Microsoft Corporation's massive software products. This shows that Microsoft Corporation prefers to invest in more AI application leaders rather than solely relying on OpenAI. Claude Code and Clawdbot/Moltbot have become popular globally. Claude Code has been extremely popular in the "AI programming/Agentic coding" community recently, as it not only "chats to help write code" but also productizes the entire engineering workflow of "writing-editing-testing-running-fixing-delivering." One of the strongest signals is revealed by Anthropic itself, as Claude Code achieved a $1 billion run-rate revenue level within approximately six months of being open to the public - a financial metric used to extrapolate annual revenue based on the company's current revenue level, commonly used to predict or assess company performance, especially for new or growing companies - something that usually only occurs with high-frequency, continuous payments from developers and enterprise customers. Claude Code is defined as an "agentic coding tool living in the terminal," naturally integrating with developers' daily surface tools like repos, CLIs, Git, and CIs, lowering migration costs. The core capability of Claude Code is "agentic" rather than "single Q&A," which has been a key factor in its rapid global popularity. Anthropic's research on SWE-bench emphasizes that to turn model capabilities into real engineering output, it is necessary to build agents around the model (task decomposition, execution, verification, rollback, etc.). Claude 3.5 Sonnet has shown leading results in SWE-bench Verified, demonstrating its "implementable repair/implementation capabilities" in both testing and practice. From a broader perspective of AI application trends, coding has become the killer scenario for generative AI to achieve "quantifiable ROI" firstly: tasks are naturally verifiable (can compile/run/test), feedback loops are short, and delivery cycles can be directly compressed. This is why the personal AI assistant ClawdBot (recently renamed Moltbot) has sparked discussions lately - it focuses not just on the "executor" aspect of IDE/terminal, but more on being a cross-application personal/team coordination agent, initiating tasks, tracking statuses, summarizing results, and even driving practices like "vibe-coding"; being open-source and self-hostable further accelerates its dissemination speed. Looking at the current AI application trend, Claude Code +(Clawdbot/Moltbot) represent the same trend: AI monetization shifting from "selling models" to "selling workflows." Coding scenarios are naturally verifiable (can compile/run/test), making it easier to establish enterprise-level ROI first; the next competition focus will shift from "single model capabilities" to "multiple agent orchestration, observability, permission openness, and security-first architecture," expanding from developer surfaces to more organizational surfaces (business systems, work orders, operations, customer service, etc.). AI programming is rapidly penetrating globally! The wave of AI applications around agent-based workflows has arrived. For the bullish narrative logic of the global stock market's AI application sector, with Anthropic being called the "OpenAI rival" raising revenue expectations, it is undoubtedly a somewhat positive growth signal - "validating the feasibility of the AI application story + preheating the potential accelerated growth trend after 2026." AI programming is currently one of the hottest technology sectors in the global AI application wave. Whether in the B2B or B2C market, agent-based "vibe coding" based on AI has become a dominating trend, accelerating its penetration worldwide. The so-called Vibe coding can be understood as a more radical and colloquial expression of AI programming (AI-assisted coding). AI programming emphasizes using the most natural human language prompts to allow AI large models to generate, modify, debug code, with the human role shifting from "writing code" to "describing goals, iterating feedback"; in some programming engineering practices, it emphasizes "running first, analyzing later," with weaker requirements for line-by-line code understanding and review. Previously, cloud computing and search engine leader Alphabet Inc. Class C, the AI application leader focused on "AI + digital advertising" Applovin, the AI software platform leader focused on "AI + data analysis" Palantir, and the "Internet gatekeeper" Cloudflare focused on "connecting clouds (Connectivity Cloud)," have all disclosed incredibly strong performance data and future performance outlook since 2025. This indicates not only the strong demand for AI computing infrastructure represented by NVIDIA Corporation AI GPU but also the strong demand for AI software applications, especially for enterprise-level AI applications that can significantly improve operational efficiency, rapidly penetrating various industries. Based on the current technological trajectory, the development direction of AI software applications is concentrated on "generative AI applications" (such as DeepSeek, ChatGPT, Sora, and Anthropic's popular AI applications like Claude) and moving forward from generative AI to AI functions transitioning from chat-box-style Q&A to "autonomous execution by agent-based AI intelligence for various tedious and complex tasks." Enterprises are urgently pushing forward the widespread application of two core categories of AI software applications - generative AI applications and AI agents - to improve efficiency and reduce operating costs. AI agents are very likely to become the major trend of AI applications before 2030. The emergence of AI agents signifies that artificial intelligence is evolving from an information assistance tool to a highly intelligent productivity tool. According to the latest research by MarketsandMarkets, the AI agent market is expected to reach $53 billion by 2030, indicating a high compound annual growth rate (CAGR) of 46% since 2025.