Anthropic’s artificial intelligence (AI) business is growing at a pace that would make most startups jealous, but investors may want to take a closer look. Roughly 20% of the Claude maker’s first-half revenue came from just two customers. That’s a pretty big slice of the pie resting in very few hands, especially for a company preparing for what could be one of the biggest IPOs in history.
The stakes get higher when we consider the costs. Anthropic spent nearly $9 billion on computing capacity and infrastructure in the first half alone, surpassing the $7.3 billion it spent throughout 2025. Meanwhile, its potential IPO could raise as much as $100 billion, although concerns over customer concentration, soaring development costs, and revenue growth have reportedly put its previously discussed $2 trillion valuation under pressure.
So, with enormous ambitions and equally enormous bills, what does this customer concentration risk mean for future Anthropic stock investors? Let’s take a closer look.
About Anthropic
Founded in 2021 and headquartered in San Francisco, California, Anthropic has quickly become one of the biggest names in the artificial intelligence industry. The company is best known for Claude, its family of AI models and chatbots, and Claude Code, which helps developers write and manage software.
But Anthropic is not just about building smarter AI tools. It also focuses on making AI systems safer, more reliable, and easier to control. Businesses have become a major source of revenue as they use Claude for coding, research, and everyday work. Anthropic has also introduced Claude Duet, expanding its efforts to bring AI into workplace tasks. As a public benefit corporation, the company aims to develop powerful AI while keeping its potential risks in check.
S-1 Filing Reveals Explosive Revenue Growth, Rising Costs, and Bigger Risks
Anthropic is growing at a speed that would make most businesses jealous, but its latest IPO filing shows that building an AI empire comes with a rather expensive bill. The company behind Claude has publicly filed its S-1 prospectus with the U.S. Securities and Exchange Commission (SEC), giving investors their first detailed look at its finances, business risks, and ambitions ahead of a potential stock market debut.
Anthropic generated $16.3 billion in revenue during the six months ended June 30, 2026, up from just $1.2 billion in the same period last year. That’s a remarkable 14x jump, reflecting growing demand for Claude and the company’s AI services. Its trailing 12-month revenue also reportedly surged approximately 1,200% between September 2025 and September 2026. Clearly, businesses are willing to spend serious money on AI, and Anthropic is getting its share.
But revenue growth does not automatically mean profits are rolling in. Anthropic reported an operating loss of -$693 million in the first half of 2026, down considerably from -$2.1 billion a year earlier. Meanwhile, its adjusted operating income improved to $1.6 billion from an adjusted operating loss of -$1.8 billion in the same period last year. That suggests the company is becoming more efficient, although it still has plenty of financial ground to cover.
Its reported net loss, however, widened to -$80.7 billion from -$13 billion in the year-ago period, according to the figures provided. A large portion reportedly came from a non-cash accounting charge linked to the increased value of its convertible securities, rather than money directly flowing out of the business. That said, don’t mistake an accounting quirk for a clean bill of financial health. Anthropic still needs huge amounts of capital to build and run its AI infrastructure, making it important for investors to look beyond the headline loss and understand the actual costs of keeping its AI ambitions alive.
Anthropic’s greatest challenge may not be attracting customers, but paying for the technology that keeps them coming back. The company spent nearly $9 billion on computing and infrastructure in the first half of 2026, compared with $2.3 billion a year earlier and $7.3 billion for all of 2025.
And the spending spree is far from over. According to its S-1 filing, Anthropic has committed between $542 billion and $567 billion toward future computing capacity and infrastructure. This includes a $417 billion deal signed in December 2025, with arrangements lasting up to 20 years, plus another $125 billion to $150 billion in additional capacity commitments. With Amazon (AMZN) and Alphabet's (GOOG) (GOOGL) Google among its computing partners, these massive obligations could put significant pressure on Anthropic’s finances if revenue growth slows.
The filing also highlighted risks beyond money. Anthropic acknowledges that its AI models could develop unexpected capabilities or cause real-world harm, potentially exposing the company to legal, financial, and reputational consequences. CEO Dario Amodei has previously warned about the dangers of advanced AI, even calling for a slower approach to developing frontier models.
What Anthropic’s Customer Concentration Means for Investors
Anthropic’s IPO prospectus revealed that roughly 20% of its first-half revenue came from just two customers, but the company did not explicitly name them. Reports and industry developments point to potential connections with platforms such as Cursor and GitHub Copilot, although the identities of the two customers disclosed in the filing have not been confirmed.
The connection makes sense. SpaceX’s (SPCX) AI business acquired Cursor, an AI coding startup whose code editor gives developers access to Anthropic’s Claude models. Meanwhile, Microsoft (MSFT) integrates Claude into GitHub Copilot, allowing developers to use Anthropic’s models alongside competing AI systems. Whenever these platforms make usage-based API requests to Claude, Anthropic can generate revenue. The more developers rely on its coding tools, the greater that revenue opportunity becomes.
But revenue concentration becomes a bigger risk when major customers can walk away without long-term commitments. According to the filing, Anthropic’s largest customers access its models on a usage-based basis without long-term contracts. If competitors such as OpenAI, Google, or xAI develop comparable coding models at lower prices, customers could shift their workloads elsewhere. Anthropic might then face slower revenue growth, pricing pressure, or both, potentially making its already expensive AI operations harder to sustain.
That does not mean these customers will leave or that Anthropic’s growth story is broken. Its competitive position in AI coding could keep attracting developers and drive demand. However, investors will need to weigh that opportunity against the possibility that a small number of customers hold considerable influence over revenue.
However, Anthropic's 335-page IPO offering document says that customer concentration has declined, with no single client contributing more than 10% of revenue in the nine months ended Sept. 30, 2026, while its top five customers accounted for roughly 25%. This marks a shift from 2025, when customers spending more than $100,000 accounted for 76% of total revenue.
Final Thoughts on Anthropic
For future Anthropic investors, the biggest takeaway is that rapid revenue growth does not automatically make a business a safe bet. Having two customers contribute roughly 20% of revenue shows strong demand, but it also leaves Anthropic exposed if either customer cuts spending or switches to a competitor.
With AI development costs soaring and no long-term contracts protecting some of its biggest customer relationships, revenue could become less predictable. Of course, Anthropic’s growing customer base and competitive AI models could help offset these risks. Ultimately, investors must decide whether its growth potential justifies the financial uncertainties.
On the date of publication, Sristi Suman Jayaswal did not have (either directly or indirectly) positions in any of the securities mentioned in this article. All information and data in this article is solely for informational purposes. For more information please view the Barchart Disclosure Policy here.