Anthropic's IPO: A close look at AI frontrunner's rapid growth, spiraling costs

By Deborah Mary Sophia

Oct. 26 (Reuters) - Anthropic filed paperwork for its planned stock market listing in the United States on Monday. This could lead to a major stock market debut. The offering will test whether investors are willing to help pay the huge costs of the AI company as concerns grow about its heavy spending, which is partly funded by debt.

The company first filed its paperwork privately in June. In recent weeks, exclusive Reuters reports have highlighted its wide-ranging plans for AI, its rising costs and concerns about the "existential risks to humanity" that advanced AI could create.

Here are some details about Anthropic's finances, how its business is organized and its position in the global AI race:

BACKGROUND AND FINANCES

Anthropic was founded in 2021 by former OpenAI researchers, including siblings Dario and Daniela Amodei. The startup makes the Claude family of AI models. Claude has four levels: Fable, Opus, Sonnet and Haiku. These models power products such as Anthropic's chatbot, a tool that helps people write computer code, and Claude Cowork, an AI agent designed for business customers.

Anthropic first focused on training its models to handle coding tasks. This helped the company become a leading competitor in the AI race and attract a steady stream of business customers.

Most of Anthropic's revenue comes from businesses that pay to use its models. The amount they pay depends on how much information the models process, measured in units called "tokens." Anthropic calls this "consumption-based" revenue. The company also earns money from paid subscriptions, including its "Pro" and "Max" plans.

This business has helped Anthropic increase its revenue from $41 million in 2023 to $386 million in 2024 and $4.59 billion in 2025. In the six months ending in June 2026, its revenue reached $16.29 billion.

But as revenue has grown quickly, losses have also increased. Anthropic reported an operating loss of $693 million in the first half of 2026. Its operating loss for all of 2025 was $8.06 billion, up from $2.98 billion in 2024.

COSTS, COMPUTING DEMAND AND AI INVESTMENTS THAT LOOP BACK

Computing power is Anthropic's biggest expense.

The company's costs include fees paid to cloud service providers such as Amazon Web Services and Google Cloud. It also pays for data center facilities and rented data center space, equipment leases and electricity.

Over the past three years, Anthropic's spending on computing and infrastructure has been much greater than its revenue.

The company has agreed to spend more than $500 billion with six partners over ten years. This is one of the largest AI infrastructure spending commitments on record. Anthropic said that many parts of these agreements cannot be canceled, or require the company to pay even if it does not use the services.

Over the next seven to ten years, Anthropic has agreed to spend at least $110 billion with Amazon, about $114.5 billion with Google and $44.9 billion with Microsoft on cloud computing.

Some of these companies are also among Anthropic's biggest investors. Amazon and Google are two of its largest shareholders. This is another example of a circular funding pattern in the AI industry, where companies invest in businesses that also spend money on their services.

Amazon has invested at least $13 billion in Anthropic and has agreed to invest up to another $20 billion. Google has committed up to $40 billion to Anthropic.

GOVERNANCE AND STRUCTURE

Anthropic is a public benefit corporation, or PBC. This means the law requires it to balance making profits for investors with developing AI that benefits humanity. An independent group chooses most of its board members.

The company is led by CEO Dario Amodei and his sister Daniela, who is president and chair of the board. Anthropic will remain a PBC after its stock market listing.

The company expects its founders and an independent trust to keep control. This setup could limit how much investors can oversee the company and how much influence ordinary shareholders have. Anthropic has said that keeping control and money in fewer hands is necessary to reduce AI risks.

Under this arrangement, a majority vote of the seven co-founders will control one share of Class F stock. This share represents 50.1% of all voting power on important company matters, including the choice of some board members and other issues put to shareholders for a vote.

COMPETITION

Anthropic's closest competitor is OpenAI, the company that created ChatGPT. ChatGPT launched in 2022 and helped kick off the worldwide race to develop and use AI technology.

After Anthropic made progress in the business market, OpenAI also increased its focus on products for software developers and business customers.

Anthropic is also competing with Elon Musk's SpaceX and Google's parent company, Alphabet. All of these companies are spending billions of dollars on AI models, data centers and special chips designed for AI as they compete to lead the field of generative AI, which creates new content.

Data from Ramp, which tracks part of the money businesses spend on AI, showed that Anthropic accounted for 60.1% of spending on AI tokens in the week ending October 11. The startup has gained ground against rivals as businesses increase their spending on AI services.

(Reporting by Deborah Sophia in Bengaluru; Editing by Arun Koyyur)