What Is Anthropic? Claude, ANT Stock and IPO

Anthropic PBC is an artificial intelligence company based in San Francisco. It builds the Claude family of large language models, treats AI safety as a core design requirement, and was founded in 2021 by former OpenAI researchers. It was valued at approximately USD 965 billion in May 2026, and on 26 October 2026 it filed to list on Nasdaq under the ticker “ANT”.

Think of Anthropic as a research lab that became one of the world’s largest software businesses in roughly five years. The company still publishes safety research and still calls itself a research company, while its legal structure obliges it to weigh public benefit alongside profit. At the same time, it reported USD 16.3 billion in revenue for the first half of 2026 and is preparing one of the largest stock market debuts on record, which helps explain why investors, AI professionals, businesses evaluating AI products, policymakers, and readers trying to understand Claude are paying close attention.

Key takeaways

• What is Anthropic? An AI research company and public benefit corporation based in San Francisco. It develops the Claude AI models: Claude Opus, Claude Sonnet and Claude Haiku.
• Valuation: A USD 65 billion Series H round in May 2026 set a post money valuation of approximately USD 965 billion.
• IPO: Anthropic filed its S-1 with the US Securities and Exchange Commission on 26 October 2026. It plans to list Class A shares on the Nasdaq Global Select Market under the Anthropic ticker “ANT”.
• Financials: Revenue for the first half of 2026 was approximately USD 16.3 billion. The net loss was approximately USD 80.7 billion, most of it a non-cash accounting charge.
• Anthropic stock price: There is no public stock price yet. Until listing, Anthropic stock is private company stock, priced only through funding rounds and secondary market activity.
• Safety focus: Constitutional AI, a responsible scaling policy and interpretability research set Anthropic apart from many other AI labs.

What is Anthropic AI?

Anthropic is an AI research and product company that builds steerable AI systems. These are models designed to follow clear instructions, explain their reasoning and stay within set limits. Its flagship product, Claude, is used by individuals, businesses and public bodies for writing, analysis, code generation, research and reasoning over long documents.

People who search “what is Anthropic AI” usually mean either the company or the Claude models it sells. Anthropic is the business. Claude is the product.

Where does the name come from?

“Anthropic” comes from the Greek anthropos, meaning “human”. The word also appears in cosmology as the anthropic principle: the idea that what we observe about the universe must allow for conscious observers to exist. The name reflects the company’s stated aim, which is AI technology that serves human values rather than working against them.

Anthropic PBC: mission and public benefit structure

Anthropic is a public benefit corporation (PBC). A public benefit corporation is a for-profit company that must, by law, balance returns to shareholders against a stated public purpose. Anthropic’s stated purpose is to develop and maintain advanced AI responsibly, for the long-term benefit of humanity.

In practice, the structure shapes three things:

• Governance: A Long-Term Benefit Trust holds special shares that let it elect part of the board. This gives a body with no financial stake a formal say in the company’s direction.
• Reporting: Anthropic publishes an annual Public Benefit Report. It tracks progress against the company’s charter goals, which range from technical AI safety to widely shared prosperity.
• Capital commitments: On 18 October 2026, Anthropic announced a Public Benefit Fund and committed 5% of its shares to it. The fund focuses on AI safety, governance and how AI affects society. The co-founders had already pledged to give away most of their own equity.

The mission rests on one position: advanced AI could carry serious, even catastrophic, risks if it is not properly aligned, so safety research has to keep pace with capability research.

Founding, history and the funding timeline

A group of former OpenAI researchers and leaders founded Anthropic in early 2021. They included Dario Amodei, Daniela Amodei, Jack Clark, Jared Kaplan, Tom Brown, Chris Olah and Sam McCandlish. The founders left OpenAI because they disagreed about how quickly powerful models should be released compared with how fast safety work could keep up.

What followed was one of the steepest funding curves in venture capital history.

Strategic partners matter as much as the cash. Amazon has invested USD 8 billion in total and is Anthropic’s main partner for cloud computing and model training. Google holds a significant minority stake and supplies computing capacity on its TPU chips.

Who has invested in Anthropic?

The later rounds drew many of the biggest names in investment management. Investors reported in the growth rounds from 2024 through the Series E to Series H rounds include:

• Venture firms: Lightspeed Venture Partners, Menlo Ventures, Bessemer Venture Partners, Founders Fund, Accel Partners, Insight Partners, General Catalyst and Sound Ventures
• Growth and crossover funds: General Atlantic, Whale Rock Capital Management, Alpha Wave Global and TowerBrook Capital Partners
• Asset managers: Capital Group, WCM Investment Management, Morgan Stanley Investment Management, Fidelity Management & Research Company and T. Rowe Price Associates
• Sovereign and pension capital: Qatar Investment Authority and the Ontario Teachers Pension Plan
• Banks and corporates: Goldman Sachs and Cisco Investments, alongside Amazon and Google

Backing from venture firms, sovereign funds and other institutional capital partners helps explain how Anthropic raised so much as a private company. It also means many existing private investors will hold shares when the company lists.

Leadership and culture

Anthropic’s leaders combine AI research backgrounds with experience in operations and finance:

• Dario Amodei (Chief Executive Officer): Former VP of Research at OpenAI. He is the company’s main public voice on strategy and AI policy.
• Daniela Amodei (President): Oversees operations, partnerships and the commercial side of the company.
• Krishna Rao (Chief Financial Officer): Leads financial strategy and the listing process.
• Jared Kaplan (Chief Science Officer): Known for early work on scaling laws, which describe how models improve as they get more data and computing power.
• Jack Clark (co-founder, Head of Policy): Leads the company’s work with governments and the policy community.

The workforce now runs into the thousands, covering machine learning, infrastructure, security, policy and emerging risk research. The company is known for testing its own models for harmful behaviour (red-teaming), holding structured safety reviews before major releases, and doing alignment work alongside every product cycle.

Core technology: how Anthropic’s models work

Anthropic’s models are large language models: neural networks trained on vast amounts of text and code to predict and generate language. Training starts with large-scale pre-training and adds reinforcement learning. One form is reinforcement learning from human feedback, where people rate the model’s answers and those ratings steer it towards responses judged helpful and safe.

Three areas of AI research stand out:

1. Interpretability. Anthropic maps the internal “features” and circuits of its models to concepts humans can understand. The aim is to inspect and fix AI systems from the inside instead of treating them as black boxes.
2. Model behaviour at scale. The company studies how increasingly capable models plan, use tools and act on their own. This scientific research feeds into its safety policies.
3. Long context and agents. Recent Claude models can work with very large amounts of text at once, up to 1 million tokens in some configurations. They can also act as agents that use tools, write code and complete tasks with several steps.

These advances in machine learning have commercial results. Agentic coding, where Claude writes, tests and edits software with little supervision, has become one of the company’s largest sources of revenue.

The Claude family: Claude Opus, Claude Sonnet and Claude Haiku

Claude is available as a web, desktop and mobile app. It is also available through an API (application programming interface), which lets developers build Claude into their own software. Anthropic is an AI research company, and anthropic’s models are offered to organizations and developers with a strong focus on alignment, safety, interpretability and AI systems guided by human values. Anthropic’s models come in three tiers.

• Claude Opus is the flagship for work where accuracy matters more than speed.
• Claude Sonnet is the workhorse. Each new Sonnet generation has tended to be a clear upgrade for business users, often matching the previous Opus on many tasks at a lower price, and Claude Sonnet 5.5 runs over 30% faster than its predecessor.
• Claude Haiku suits real-time products and very large numbers of requests.

API use is charged per million tokens and the price varies by model. A token is a small chunk of text, roughly three quarters of a word. Anthropic publishes current rates on its own pricing page, and they change as new versions launch.

Constitutional AI and Anthropic’s approach to AI safety

At Anthropic, AI safety is built into training rather than added as a filter afterwards. The best-known technique is Constitutional AI.

Constitutional AI is a training method in which a model learns to judge and improve its own answers against a written set of principles, called a constitution. The constitution draws on ethical standards, human rights principles and Anthropic’s own guidelines. It sets out how Claude should weigh being safe, being ethical and being helpful.

Three other safeguards sit alongside it:

• Responsible Scaling Policy: A framework that requires tighter security and release controls as models become more capable. Anthropic publishes assessments covering areas such as cybersecurity and misuse.
• Risk reporting: Anthropic publishes risk reports that explain how it assesses and reduces specific harms.
• Red-teaming: Internal and external testers try to make models behave dangerously before they are released.

Critics argue that no lab can fully prove its most advanced systems are safe. Supporters point out that Anthropic publishes more of its methods than most rivals. Either way, safety is central to how the company presents itself to customers, regulators and future shareholders.

Products and real-world use cases

Anthropic sells mainly to organisations, although millions of individuals use Claude directly.

• Software development: Claude Code and the API are used to generate, review and restructure code. Reporting on the S-1 suggests coding tools alone brought in billions of dollars of annualised revenue in 2026.
• Enterprise knowledge work: Summarising contracts, drafting reports, analysing spreadsheets and building internal search tools.
• Financial services: Banks and asset managers use Claude for research summaries, document review and analysis. Some of the same institutions are also Anthropic shareholders.
• Science and health: Partnerships in drug discovery and other areas of scientific research.
• Education and public sector: Agreements with universities and government departments to adopt AI.

Government, national security and policy

Anthropic works with the US government and other governments on testing AI, setting safety standards and deploying AI responsibly. It has made special versions of Claude, with extra security controls, available for national security and defence work.

That relationship is not always smooth. Anthropic’s usage policies restrict some military and surveillance uses, which has caused friction with agencies that want fewer limits. This matters to investors because government contracts are a growing source of revenue, and policy disputes can affect them.

Can Claude be used as a trading bot?

Searches for a “Claude trading bot” have grown along with interest in AI-driven trading. In short, Claude is a general-purpose AI assistant, not a trading bot, and Anthropic does not sell an automated trading product.

Traders can, however, use Claude as one tool in their own process:

• Research support: Summarising earnings reports, central bank statements or economic data releases.
• Code generation: Drafting or debugging scripts for platforms such as MetaTrader, including Expert Advisors (EAs), which are automated trading programs.
• Strategy review: Explaining technical indicators or helping to set up a backtest, which tests a strategy on past price data.

Keep a few precautions in mind:

• Take note: Large language models can give confident but wrong answers, including incorrect figures or faulty code. Check every output.
• Reminder: An AI model has no live market data unless it is connected to a data source, and it cannot predict prices.
• Caution: Test any automated strategy built with AI help on a demo account before using real money.

The guide to the best AI stock trading bots compares how dedicated trading tools work.

Anthropic stock: IPO filing and what the S-1 shows

As of early November 2026, Anthropic is still a private company, but it is well on the way to listing on public stock exchanges.

• Filing: Anthropic submitted a confidential draft in mid-2026. It then filed its public S-1 registration statement with the SEC on 26 October 2026, as reported by the Associated Press.
• Exchange and ticker: The company plans to list Class A common stock on the Nasdaq Global Select Market under the Anthropic ticker “ANT”.
• Underwriters: Goldman Sachs and Morgan Stanley lead a large group of banks managing the offering.
• Share structure: According to reports on the filing, the listed Class A shares carry one vote each. Separate share classes give the Long-Term Benefit Trust and the co-founders significant control.
• Timing: Media reports point to a listing in November 2026. The company has not confirmed a date, a share count or a price range.

Key numbers from the filing

The headline loss looks alarming, but most of it was a one-time accounting adjustment linked to financing arrangements, not cash spent running the business. Even so, Anthropic spends heavily on computing power, and its long-term infrastructure commitments run into the hundreds of billions of dollars.

Anthropic stock price: how is it valued before listing?

Anthropic has no official stock price until its shares start trading on Nasdaq. Until then, its value is estimated in two ways.

Funding rounds. Each round sets a post money valuation: the company’s value immediately after the new money comes in. The Series H figure of approximately USD 965 billion is the most recent valuation backed by the company.

Secondary market activity. Employees and early investors sometimes sell shares to other buyers on private market trading platforms. Firms such as Forge Global publish Forge Price and Forge Data. These indicators are built from secondary trades and funding data, and aim to give deep data insights into the implied stock price of a specific company. Some shareholders worldwide rely on these figures as a reference point. However, they come from a small, restricted market and can differ widely from where a stock eventually trades.

Reports ahead of the listing have suggested a target valuation between USD 1.5 trillion and USD 2 trillion. These are media estimates, not company guidance. The final price will depend on investor demand during the roadshow, when the company and its banks market the shares to investors.

Where can I buy Anthropic stock?

The answer depends on timing and on what kind of investor you are.

Before the IPO

• Secondary marketplaces: Some platforms offer liquidity solutions and tailored investment opportunities in private company stock. These are usually limited to accredited or professional investors and have high minimum investments.
• Private funds: Some venture or pre-IPO funds hold Anthropic shares indirectly.
• Restrictions apply: Shares in private companies often have transfer limits and need company approval to sell, so they are hard to come by.

No platform can offer guaranteed returns on pre-IPO shares. A marketplace is not a trusted platform just because it lists a famous name. Take note of fees, how the valuation is worked out and how easily you could sell a position.

After the IPO

Once Anthropic lists, ANT shares should be available through stockbrokers with access to Nasdaq, as with other public companies. Some brokers may also offer contracts for difference (CFDs) on the stock. A CFD is a contract that lets you speculate on a share’s price movement without owning the share. Whether a product is offered on a new listing varies by broker and region.

Traders who want broader exposure to large US technology companies sometimes look at index products instead. The Nasdaq 100 guide explains how that index works and how newly listed companies can eventually join it.

Can you sell Anthropic stock before it lists?

Before the IPO, existing holders who want to sell Anthropic stock generally need company approval and a buyer found through an approved process, such as a tender offer or an approved secondary sale. After listing, insiders are usually bound by a lock-up period, often around 180 days, during which they cannot sell freely. A wave of selling when a lock-up ends often puts pressure on the price of newly listed shares.

Precautions to consider before investing in an AI IPO

Large technology IPOs attract a lot of attention, and early trading can be volatile. Keep a few points in mind:

• Valuation reminder: At a reported USD 1.5 trillion or more, the price would assume years of rapid growth.
• Cost caution: Computing power, chips and data centres are expensive, and spending commitments are large compared with current revenue.
• Competition: OpenAI, Google DeepMind, Meta and developers of open-weight models compete on price and capability, and gaps in model performance can close quickly.
• Regulation: AI rules in the US, the EU and elsewhere are still being written.
• Legal matters: Like its peers, Anthropic has faced copyright lawsuits over its training data, and the outcomes could affect costs and how models are built.
• Governance: Because the founders and the trust hold strong control rights, public shareholders may have little say in major decisions.

For comparison with another AI leader, the OpenAI stock guide covers ownership and access before an IPO in a similar way.

How Anthropic compares with other AI labs

The biggest practical difference is who the customers are. Anthropic earns most of its revenue from businesses and developers, while OpenAI has a much larger consumer base. Anthropic is betting that businesses will pay more for reliable, steerable AI systems with strong safety controls.

Start online CFD trading with VT Markets today

If you want to follow how AI companies and the indices that hold them move, VT Markets gives you trading tools and platforms to get started. You can trade on MetaTrader 4 (MT4) and MetaTrader 5 (MT5), which offer fast execution, advanced charting and support for automated strategies, including Expert Advisors built with AI help. CFDs on US stocks and major indices let you trade on price movements without owning the underlying asset. Leverage magnifies both gains and losses.

New to trading? Practise with virtual funds on a VT Markets demo account before you start trading with real money.

Open your account with VT Markets today and access secure, transparent, and competitive CFD trading across some of the world’s most popular markets.

Frequently asked questions

Is Anthropic an AI or a company?

Anthropic is a company run by people, not an AI system. When people say “Anthropic is an AI”, they mean it is an AI research company. Its AI models are called Claude, and the company employs thousands of researchers, engineers and policy specialists.

What is the Anthropic ticker symbol?

Anthropic has applied to list its Class A common stock on the Nasdaq Global Select Market under the ticker “ANT”. The symbol becomes active only once the IPO is priced and the shares start trading.

How does Anthropic make money?

Most revenue comes from three sources: usage-based API fees paid by developers and businesses, enterprise contracts, and Claude subscriptions for individuals and teams. Amazon and Google also offer Claude to their cloud customers. Revenue in the first half of 2026 was approximately USD 16.3 billion.

Can individuals use Claude?

Yes. Claude is available on the web, desktop and mobile, with a free tier and paid plans. Developers and organisations can also use it through the API and enterprise agreements.

Risk Warning: Trading Contracts for Difference (CFDs) carries a high level of risk and may not be suitable for all investors. The use of leverage can significantly magnify gains and losses and may result in losses exceeding your initial investment. Prior to engaging in CFD trading, you should ensure that you fully understand the risks involved, carefully consider your investment objectives, financial situation, and level of experience, and seek independent advice where necessary. Past performance is not indicative of future results. Please refer to our legal documents for a comprehensive overview of the risks associated with CFD trading.

General Disclaimer: The content on this website is provided for general informational purposes only and does not take into account your specific investment objectives, financial circumstances, or particular needs. Access to this website is made at your own initiative.