Anthropic’s AI coding assistant, Claude Code, is emerging as one of the company’s biggest commercial successes, with new financial figures revealing impressive customer spending growth and a surprising improvement in profitability.
The artificial intelligence company reportedly recorded an operating profit in the second quarter, while its net revenue retention reached an extraordinary 900%. The figure means that spending by the same group of customers in the third quarter of 2026 was nine times higher than their spending in the corresponding quarter of 2025.
The numbers highlight how successfully Anthropic has attracted businesses to its AI-powered coding tools and encouraged existing customers to increase their usage.
However, rapid adoption is creating another challenge: companies increasingly want to understand whether the money they spend on artificial intelligence is actually delivering measurable business benefits.
In response, Anthropic has introduced a new collection of reporting and spending management tools designed to help enterprise customers monitor their AI usage, evaluate productivity and identify opportunities to reduce unnecessary costs.
The move comes as businesses face growing pressure to demonstrate returns on their investments in AI, particularly as spending on automated coding tools continues to rise.
Claude Code Drives Anthropic’s Growing Enterprise Business
Anthropic has increasingly focused on winning enterprise customers, and Claude Code has become an important part of that strategy.
The AI coding assistant helps software developers write, analyse, modify and manage computer code, allowing businesses to automate parts of their software development processes.
Its growing popularity has helped Anthropic establish a strong position in the market for AI-powered programming tools.
The company’s latest financial figures suggest that existing customers are not only continuing to use its products but are also spending substantially more on them.
Its reported 900% net revenue retention rate is particularly significant because it indicates that spending from existing customers has expanded dramatically over the past year.
For a company operating in the highly competitive AI industry, retaining customers and increasing their spending can be just as important as attracting new users.
Anthropic appears to be building its business around this principle by encouraging organisations to integrate Claude Code into their everyday operations.
Once companies begin relying on an AI coding assistant for important development tasks, the product can become deeply embedded in their engineering processes.
This creates opportunities for Anthropic to generate more revenue from customers as their use of AI expands.
Anthropic Introduces New Tools to Help Businesses Track AI Spending
Despite the impressive adoption figures, businesses are becoming increasingly concerned about the cost of using advanced AI systems.
Many organisations initially embraced AI tools because of their potential to improve productivity, reduce repetitive work and accelerate software development.
However, as usage increases, so do the associated expenses.
AI services often charge customers based on token consumption, meaning businesses can accumulate significant costs when employees and automated agents use these systems extensively.
Anthropic’s latest reporting tools are designed to help organisations understand exactly where their money is going and whether the resulting work justifies the expense.
Announced on Wednesday, the new features provide enterprise managers with more detailed information about how their teams use Anthropic’s AI products.
Using internal analytics and employee surveys, managers can examine team activities over selected periods and evaluate several aspects of their AI-assisted work.
The reports can show:
- What work individual teams have completed.
- The amount of money spent on AI usage.
- The output produced by employees and AI tools.
- Areas where existing workflows encounter problems.
- Recommendations for improving productivity and results.
This information is intended to give businesses a clearer picture of the relationship between AI spending and actual work delivered.
Anthropic says the reporting data will not be used for model training by default, an important consideration for organisations concerned about protecting internal business information.
The company is also introducing pooled group budgets, allowing teams to manage their token spending more effectively.
Together, these features represent an effort to make AI expenditure more transparent and easier for businesses to control.
Why Companies Are Becoming More Concerned About AI Costs
The growing demand for AI reporting tools reflects a wider challenge facing businesses.
Although AI adoption has accelerated, many large organisations are still trying to determine how best to integrate the technology into their operations.
Startups backed by venture capital have been particularly enthusiastic about experimenting with AI tools, contributing to rising token consumption.
However, many traditional businesses, including companies within the Fortune 500, have yet to experience major productivity improvements from their AI investments.
For executives and financial officers, this creates an important dilemma.
They are under pressure to invest in artificial intelligence to remain competitive, but they also need evidence that the technology is generating enough value to justify its costs.
Simply allowing employees to use more AI tools does not necessarily translate into higher productivity or improved profitability.
Companies must understand which applications are genuinely useful, where AI saves time and where unnecessary spending is taking place.
Anthropic’s reporting features appear designed to address precisely these concerns.
Rather than focusing exclusively on selling more AI usage, the company is attempting to help customers make better use of the services they already purchase.
This could become an important competitive advantage as businesses become more selective about their technology budgets.
AI Coding Tool Spending Has Increased 10 Times, New Report Finds
New research from Jellyfish, a company specialising in software development management tools, provides further evidence of the rapid expansion of AI coding expenses.
The company’s latest report examined the use of agentic coding tools among 276,000 of its customers.
Agentic coding tools are AI systems capable of performing more complex programming tasks with varying degrees of independence, rather than simply suggesting individual lines of code.
According to Jellyfish, Claude Code and other tools powered by Anthropic models remain among the most widely used solutions in the survey.
Together, these tools were responsible for nearly one-third of all proposed code changes submitted for review.
Even more striking was the increase in spending.
Jellyfish found that expenditure on agentic coding tools had increased approximately tenfold since the beginning of the year.
Their costs now account for roughly 2% of total engineering expenditure.
Although that percentage might initially appear relatively small, the speed at which spending has increased is attracting attention from financial executives.
Nicholas Arcolano, Jellyfish’s chief scientist, told TechCrunch that the rapid growth was something many chief financial officers had not anticipated.
He explained that businesses were reaching a point where simply spending money on AI was no longer sufficient.
Organisations now needed to optimise their usage and ensure the technology delivered meaningful results.
His comments highlight why companies are increasingly interested in tools that can measure AI productivity and identify unnecessary expenditure.
Why Anthropic Is Keeping Its Lead Despite OpenAI’s Competition
Anthropic’s growing enterprise presence comes as competition in AI-powered coding becomes increasingly intense.
OpenAI has continued developing competing products, including its Astra model, which appeared to present a potential challenge to Anthropic’s position.
However, data from Jellyfish suggests that OpenAI’s competitive momentum has not yet significantly disrupted Anthropic’s adoption among enterprise customers.
Nikolas Albarran, Jellyfish’s AI product lead, explained that Codex took considerable time to become noticeably prominent in the company’s usage data.
Its adoption subsequently increased following the arrival of Astra, OpenAI’s latest model.
Nevertheless, Anthropic’s products continued to maintain a strong position.
Albarran noted that discussions on social media, particularly among technology enthusiasts in Silicon Valley, often create the impression that competition between AI coding products is more intense than actual enterprise adoption figures suggest.
He explained that Anthropic had established an advantage by building strong relationships with businesses and providing products that worked reliably within their existing development environments.
Many organisations had already invested considerable time and effort in standardising their software development workflows around Claude Code.
As a result, switching to another AI coding platform would require a convincing reason.
Even if a competing product offers certain advantages, businesses may be reluctant to abandon a system that their employees already understand and use successfully.
Why Businesses Are Finding It Harder to Switch AI Providers
One of the biggest advantages Anthropic appears to enjoy is the difficulty of changing AI providers once a company has integrated a particular platform into its operations.
For individual programmers, switching between coding assistants can be relatively straightforward.
They can experiment with different tools, compare their performance and choose whichever product best suits their needs.
Large organisations face a different situation.
Enterprise customers often spend substantial amounts of time integrating AI systems into internal processes, training employees and establishing standardised workflows.
They may also need to modify development procedures, configure automated agents and adapt internal documentation to work with a particular platform.
These investments create switching costs.
Even when competing products offer similar capabilities, differences in how they handle configurations, instructions and development environments can make migration complicated.
Albarran pointed specifically to differences in how AI coding tools manage files such as agent.md, which can contain instructions and configuration information used by coding agents.
Such differences may appear minor, but they can become important when an organisation has hundreds or thousands of developers relying on established systems.
The result is that a company may remain with its existing provider even when alternative products become increasingly competitive.
For Anthropic, this creates a potentially valuable long-term advantage.
Anthropic’s New Reporting Tools Could Reduce the Need for Expensive AI Consultants
Another important factor in enterprise customer retention is the growing investment major AI companies are making in forward-deployed engineering teams.
These teams typically consist of technical specialists and AI consultants who work directly with customer organisations.
Their role is to help businesses implement AI systems, customise workflows, solve technical problems and get their employees using the technology effectively.
This hands-on support can be particularly valuable for large organisations that lack the internal expertise required to deploy advanced AI tools.
However, providing this level of assistance requires significant resources.
Anthropic’s new reporting features could offer a lighter version of that support.
By giving customers detailed information about their AI usage, identifying workflow problems and recommending improvements, the company can help businesses optimise their operations without requiring the same level of direct intervention.
Albarran described the situation as an indication that enterprise AI adoption is reaching a more mature stage.
He argued that once organisations have invested substantial effort in adopting a particular platform, switching becomes increasingly difficult.
Even when competing products are broadly similar, differences in their technical behaviour and workflows mean that changing providers is not entirely cost-free.
This suggests that enterprise success in the AI industry will depend on more than developing powerful models.
Customer support, integration, reliability and the ability to demonstrate measurable business value are becoming equally important.
The Future of AI Coding Will Be About Results, Not Just Spending
Anthropic’s latest developments reveal an important shift in the artificial intelligence industry.
During the early stages of the AI boom, companies were primarily focused on attracting users, improving models and expanding access to new capabilities.
Now, as businesses spend increasingly large amounts on AI services, the conversation is changing.
Executives want to know whether these investments are improving productivity, reducing costs and generating measurable returns.
Jellyfish’s findings suggest that AI coding expenditure is growing rapidly, while Anthropic’s new reporting tools demonstrate that customers are demanding better ways to monitor that spending.
For Anthropic, the strategy could strengthen its position among enterprise customers by making Claude Code more valuable and harder to replace.
For businesses, however, the priority is ensuring that increased AI adoption produces tangible improvements rather than simply creating another growing expense.
The next phase of competition between Anthropic, OpenAI and other AI developers may therefore depend less on which company produces the most impressive demonstrations and more on which one helps businesses achieve the best results.
As AI coding tools become a regular part of software development, the companies that can combine powerful technology with reliable workflows, cost management and measurable productivity gains are likely to have the strongest advantage.
