Anthropic and OpenAI Grab 86% of AI API Market

Anthropic and OpenAI together account for 86% of enterprise large language model (LLM) API spending, highlighting how two companies are consolidating their positions in the rapidly expanding business AI market. According to venture capital firm Menlo Ventures’ 2026 State of Enterprise AI report, released on November 3, Anthropic holds 55% of enterprise LLM API spending, while OpenAI accounts for 31%. The findings suggest that businesses are increasingly directing their AI budgets towards a small group of leading model developers as companies move from experimentation to large-scale deployment.

The market-share figures arrive as Anthropic prepares for a potential stock market listing. The company filed paperwork for a possible initial public offering on October 26, according to the Associated Press, and is reportedly targeting a Nasdaq debut under the ticker symbol ANT. The potential listing could become one of the largest IPOs in history, placing Anthropic’s business growth, financial performance and infrastructure spending under greater public scrutiny. Together, the developments underline the growing commercial importance of AI models and the competition between leading companies to capture enterprise customers.

Anthropic Leads OpenAI in Enterprise AI API Spending

Menlo Ventures’ latest report shows that Anthropic has strengthened its position in the enterprise LLM API market over the past year.

Anthropic’s share increased from 40% in 2025 to 55% in 2026. OpenAI also expanded its position, rising from 27% to 31%. Together, the companies increased their combined share from 67% to 86%.

Google experienced the biggest decline among the leading providers, with its share falling from 21% to 9%.

AI model provider	2025 market share	2026 market share
Anthropic	40%	55%
OpenAI	27%	31%
Google	21%	9%
Other providers	12%	5%
Anthropic + OpenAI	67%	86%

Figures represent estimated shares of enterprise LLM API spending reported by Menlo Ventures, not overall AI industry revenue.

Anthropic’s growth reflects rising enterprise demand for its Claude models, particularly in software development, coding assistance, reasoning and business automation.

OpenAI has maintained a substantial position through its GPT models and developer API ecosystem. Its products are used by businesses to build customer support systems, internal assistants, research tools and AI-powered applications.

The report suggests that both companies are benefiting from the increasing use of advanced AI models in production environments.

Enterprise AI Spending Reaches $176.8 Billion

The concentration of API spending is taking place alongside rapid growth in the broader enterprise AI market.

Menlo Ventures estimates that enterprise AI spending reached $176.8 billion in 2026, representing a 5.4-fold increase in one year.

Businesses are allocating money to a range of AI applications, including model access, infrastructure, software development, automation and AI-powered enterprise tools.

The report also highlights a shift in how companies use AI. Rather than limiting deployment to employees asking questions through chatbots, businesses are increasingly building systems that can complete multi-step tasks with limited human intervention.

These systems, commonly called AI agents, can interact with software, retrieve information, write code, process documents and execute parts of business workflows.

The growing commercial use of agents is creating additional demand for model APIs because these systems can make repeated model calls while completing a single task.

AI Agents Become Major Consumers of Model Tokens

One of the notable findings in Menlo Ventures’ report concerns the increasing share of AI usage generated by autonomous agents.

According to the report’s OpenRouter data, AI agents now account for approximately 80% of tokens processed on the platform, generating around 130.9 trillion tokens per week.

This represents a major shift from traditional chatbot usage, where people directly submit prompts and receive responses.

AI agents can perform several operations in sequence. For example, a software development agent might inspect a codebase, identify an error, modify files, run tests and revise its work. Each operation can require additional model interactions.

The growing number of these interactions increases demand for API access and computing infrastructure.

However, OpenRouter’s figures represent activity on its platform and should not be treated as a measurement of all global AI usage.

The broader trend is nevertheless important for AI developers because agent-based applications can generate substantial model consumption even when human users are not continuously interacting with the system.

Why Businesses Are Choosing Anthropic and OpenAI

The growing share of Anthropic and OpenAI reflects the importance businesses place on model performance, reliability and integration.

Enterprise customers often need AI systems that can handle complex instructions, process large amounts of information, generate code and connect with existing business software.

Anthropic has built a strong position around Claude, which has gained adoption in software engineering, enterprise research and AI-agent workflows.

OpenAI, meanwhile, has a broad developer ecosystem supported by its GPT models, API products and partnerships with cloud providers and enterprise software companies.

For businesses, selecting an AI model is not simply about comparing subscription prices. Accuracy, response speed, security, data handling, reliability and integration costs can all influence purchasing decisions.

The concentration also presents opportunities for smaller model developers. Companies such as Google, DeepSeek and other open-model providers can compete through lower costs, specialised capabilities and greater flexibility for organisations that want to operate models independently.

Anthropic’s IPO Could Become a Landmark AI Listing

Anthropic’s public-market ambitions are developing as demand for its AI models expands.

The Associated Press reported that Anthropic filed paperwork on October 26, 2026, for a potential IPO. The company is reportedly considering a Nasdaq listing under the ticker symbol ANT.

Reports have suggested that the offering could raise as much as $100 billion and value the company at approximately $1.5 trillion to $2 trillion. These are reported possibilities, not confirmed offering terms or a final valuation.

A listing would give public-market investors an opportunity to invest directly in one of the leading developers of frontier AI models.

It would also bring greater attention to the company’s financial statements, infrastructure commitments, customer concentration and ability to convert revenue growth into sustainable profitability.

Anthropic has been expanding its AI infrastructure through major partnerships with cloud providers and computing companies. Training and operating advanced models requires substantial investment in specialised chips, data centres, electricity and engineering.

The IPO could provide capital to support these requirements, although the company’s final fundraising plans have not been confirmed.

Revenue Growth Comes With Significant Financial Costs

Anthropic’s rapid commercial expansion has been accompanied by substantial financial commitments.

The Associated Press reported that the company generated more than $16 billion in revenue during the first half of 2026. It also reported an $80.7 billion net loss, a figure heavily affected by a one-time accounting adjustment.

The reported loss should therefore not be interpreted as a straightforward measure of the company’s recurring operating performance.

Nevertheless, the financial figures highlight the unusual economics of the AI industry. Companies developing frontier models must spend heavily on computing infrastructure, research, talent and product development before the long-term profitability of their businesses is established.

Anthropic’s IPO will give investors an opportunity to examine these costs more closely.

Its relationship with major cloud providers is another important consideration. Amazon and Google are both investors, infrastructure partners and distribution channels for Anthropic. Such relationships provide access to computing capacity and enterprise customers but also create dependencies that investors will need to evaluate.

Open-Source AI Models Face a Different Competitive Opportunity

Although Anthropic and OpenAI dominate enterprise API spending, open-source and open-weight models continue to attract significant developer interest.

Menlo Ventures estimates that open-source models account for approximately one-fifth of LLM tokens but only around one in 25 dollars of model spending.

This difference suggests that usage volume and commercial spending are not necessarily moving together.

Open-weight models can be downloaded and deployed on an organisation’s own infrastructure, giving businesses more control over data, customisation and deployment costs.

However, self-hosting can require significant investment in hardware, maintenance, security and technical expertise.

For companies that want managed services and access to leading proprietary models, commercial APIs can remain a more convenient option.

The market therefore appears to be developing along two parallel paths: large commercial providers competing for enterprise spending and open-model developers offering alternatives for organisations prioritising flexibility and control.

The Bigger Picture

The 86% combined market share of Anthropic and OpenAI illustrates how quickly the enterprise AI market is concentrating around a small number of leading model providers. As businesses move towards AI agents and automated workflows, the value of model access is expanding beyond conventional chatbot applications.

However, market leadership does not guarantee long-term dominance. Google, open-model developers and emerging AI companies can challenge existing providers through improved performance, lower prices and specialised products. The next stage of competition will depend on whether leading AI companies can maintain technical advantages while controlling the enormous costs of developing and operating advanced models.

Looking Ahead

Anthropic’s potential IPO could become a major test of investor appetite for AI companies. Investors will examine its revenue growth, financial commitments, infrastructure partnerships and ability to generate sustainable profits. The proposed listing could also influence how public markets value other private AI businesses, including OpenAI.

For the enterprise AI market, the central question is whether Anthropic and OpenAI can maintain their combined 86% spending share as competitors improve their models and businesses gain more deployment options. The growing use of AI agents is likely to increase demand for model APIs, but pricing pressure and infrastructure costs will remain important factors shaping the industry’s future.