Anthropic's initial public offering prospectus reveals that $2.16 billion of its 2025 revenue came from sales through Amazon and Google. This disclosure highlights the company's deep operational dependence on major technology partners as it prepares for a listing valued above $2 trillion.

The filing, seen by Reuters and Benzinga, underscores the concentration risk inherent in Anthropic's distribution model. While the company generated nearly $4.6 billion in total revenue in 2025, the significant portion routed through Amazon Web Services and Google Cloud indicates that these partnerships are critical to its commercial reach. The document further notes that two unnamed customers each accounted for 12% of last year's revenue, raising concerns about customer dependency.

Financial context and IPO plans

The technical issues arise as Anthropic reportedly prepares for an IPO targeting a valuation above $2 trillion. The company generated nearly $4.6 billion in revenue in 2025 but recorded a net loss of $42 billion. It plans roughly $518 billion in future cloud, computing and infrastructure spending. As of December 31, the company held $20.28 billion in cash, cash equivalents and short-term investments.

Robert Kindler, global chair of the M&A Group at Paul Weiss, does not expect the listing to move the broader stock market. Asked by CNBC how important the IPO would be, Kindler said, "I don't think it matters." He noted that companies can now raise large amounts of private capital, which reduces the need to go public to fund growth. Conversely, tech analyst Dan Ives has called the potential listing a watershed event that could benefit the entire technology sector by bringing greater transparency to the AI industry.

Revenue composition and cloud reliance

Anthropic's revenue structure is heavily weighted toward usage-based pricing. Of the total revenue, $3.8 billion came from customers paying based on their usage of the Claude AI system, while subscription revenue amounted to $789 million. The company anticipates consumption-based revenue to constitute "the substantial majority" of its revenue in the future.

Cloud marketplace sales generated roughly $2.16 billion, accounting for 47% of Anthropic’s 2025 annual revenue. By early 2026, its long-term commitments had surpassed $417 billion, covering 3.5 gigawatts of dedicated computing capacity. Despite the risks, Anthropic views these relationships as beneficial, stating that it can leverage the vast sales networks of Amazon, Google, and Microsoft Corp.'s cloud platforms to reach customers and accelerate market penetration.

What the numbers show

The new data clarifies the structure behind Anthropic's reported financials. With total 2025 revenue at $4.6 billion, the $2.16 billion attributed to Amazon and Google represents approximately 47% of the company's top line. This concentration suggests that nearly half of Anthropic's commercial success is tied to the distribution channels of two specific tech giants, rather than direct enterprise sales or other channels.

Furthermore, the disclosure that two unnamed customers each account for 12% of revenue means that just two clients drive nearly a quarter of the company's income. Combined with the reliance on Amazon and Google, this creates a significant single-point-of-failure risk if any major partner or client alters their spending patterns, especially since many lack long-term contracts.

Global impact of the disruption

The service interruption affected users internationally over the preceding 24 hours. Downdetector recorded significant report volumes across several key markets.

Anthropic's Claude AI chatbot suffered a widespread service disruption on Tuesday morning, with full service restored within approximately 40 minutes. User-reported outages on Downdetector peaked at 10,105 in the U.S. around 10:30 am ET. Anthropic's status page indicated elevated errors across claude.ai, Claude Code, Claude Cowork and the Claude API starting at about 10:21 am ET. The company confirmed the issue was fully resolved by 10:59 am ET. Users encountered failed requests, conversations that would not load and login problems.

Industry safety debates

The outage coincides with ongoing industry discussions on AI safety. OpenAI recently canceled its GPT-6.1 Astra release due to internal safety standards. Anthropic CEO Dario Amodei has advocated for safety measures keeping pace with technological advances.

President Donald Trump met with Amodei on Sunday to discuss AI development. Speaking to Fox News, Trump said he opposes slowing the pace of AI progress, citing competition with China. Trump is also scheduled to meet at the White House on Tuesday with executives from Alphabet Inc., Microsoft Corp. and Nvidia Corp.

Anthropic hired Ben Hawkes, a longtime Google security researcher and former Project Zero lead, to direct the cybersecurity mission of its Frontier Red Team. This appointment adds a veteran vulnerability researcher as AI systems increasingly demonstrate capability in sophisticated cyber tasks.

Hawkes announced the move on X, stating he would work alongside Logan Graham, the head of the Frontier Red Team. His primary objective is to address fundamental insecurities in technology that have persisted for four decades. "What will it take to make the past 40 years of hacking a historical oddity? That's what I'm here to find out," Hawkes wrote.

Background and previous roles

Before joining Anthropic, Hawkes spent more than 12 years at Google. He served nearly seven years as a lead for Project Zero, the company's security research team known for uncovering vulnerabilities in widely used software and hardware. He later served as an ISE Systems Team Lead before leaving Google in September 2026. Hawkes also founded the security research company Isosceles.

The hiring aligns with Anthropic's expansion of red-team research into areas including cryptography, autonomous systems, intelligence targeting, and weapons development. The company has reported rapid improvements in its models' ability to perform cybersecurity tasks. In a 2025 assessment, Anthropic noted that Claude had progressed significantly on capture-the-flag challenges involving software vulnerabilities.

Broader executive expansion

Anthropic has been actively building out its roster with professionals across the artificial intelligence and technology spectrum. Recent hires include:

• Matt Clifford: Hired earlier this month as Managing Director of International Affairs to lead engagement with governments across Europe, Asia-Pacific, and beyond.
• Robert Mahari: Appointed in August as the company's first head of Claude for Legal, working alongside Mark Pike to shape legal AI offerings.
• Andrej Karpathy: Hired in May; a co-founder of OpenAI and former director of artificial intelligence and Autopilot Vision at Tesla.

IPO context

Anthropic is targeting an initial public offering that could rival or exceed the record set by SpaceX (NASDAQ: SPCX). SpaceX targeted $75 billion for its IPO and ultimately raised about $86.2 billion once the overallotment option was exercised. Anthropic aims to match or exceed that figure, reflecting significant investor enthusiasm around the artificial intelligence industry.