AI Infrastructure & Markets

Amazon and Google Will Together Own Nearly 30% of Anthropic After Its IPO

What Happened

Anthropic’s planned IPO is also a major restructuring of its strategic relationships. According to TechCrunch, Amazon is expected to hold a 14.7% stake after the offering, while Google is expected to hold 13.7%. Together, the two infrastructure providers would own nearly 30% of the AI company.

Those holdings are being created through conversions of investment notes into a special class of non-voting shares. Amazon converted part of its notes at prices of approximately $2.09 or $5.05 per share, depending on the year of investment. It is retaining billions of dollars in notes; if all of them were converted immediately, Amazon’s stake would rise to approximately 23.8%.

Google’s position also combines note conversions with a private-placement share purchase. The result is an ownership structure that ties Anthropic closely to two companies that are also central to its computing strategy.

Why the Infrastructure Relationship Matters

Anthropic has committed to spend more than $100 billion on AWS cloud services over the next decade, according to TechCrunch. That commitment gives the ownership story an operational dimension: Amazon is not only an investor, but also a major infrastructure provider.

Anthropic’s compute strategy is broader than a single cloud or chip supplier. Fortune reported that the company relies on four main chip providers: Google TPUs, AMD GPUs, Amazon Trainium GPUs and Nvidia GPUs. Anthropic has also built a proprietary orchestration layer to direct workloads across different chips and reduce costs.

For platform engineers, this is the important systems detail. The company’s infrastructure is being organized around workload placement: different accelerators can be used for different jobs, while an orchestration layer helps manage the complexity. The approach reflects a wider industry effort to reduce dependence on Nvidia and diversify AI compute.

The Anthropic IPO Playbook: Follow the Signal

Amazon + Google: Nearly 30%

Amazon is expected to hold 14.7% and Google 13.7% after the IPO. Both positions are in non-voting shares created through note conversions, with Google’s position also including a private-placement purchase.

A Large Private-Placement Signal

AMD, Broadcom, Google, Micron, Microsoft, Nvidia and TSMC Partners agreed to buy $26 billion of the non-voting stock. Google and Nvidia each agreed to buy $10 billion. CNBC reported that Anthropic was seeking as much as $64.4 billion in the IPO and $26 billion through the private placement.

Infrastructure Diversification

Anthropic uses Google TPUs, AMD GPUs, Amazon Trainium GPUs and Nvidia GPUs. Its orchestration layer directs workloads among those providers, creating a technical strategy built around flexibility and cost management.

The IPO Is Also a Market Test

CNBC reported that Anthropic was likely to price above its initial $70-to-$78 range. The company planned to close its order books at 4 p.m. ET on November 17, list on Nasdaq under the ticker ANT and begin trading on November 19.

The prospectus figures explain why investor attention is so high. Anthropic reported $33.24 billion in revenue for the nine months ended September 30, up approximately 1,200% year over year. It also reported a $104 billion net loss for that period, largely attributed to changes in the fair value of convertible notes and derivative liabilities. CNBC reported that growth had been driven substantially by Claude Code, Anthropic’s AI coding tool.

Another CNBC report described the offering as a major test for the AI startup ecosystem and venture-capital markets. Anthropic was reportedly targeting a valuation of approximately $1.4 trillion, while AI startups received $355.9 billion, or 86% of U.S. venture-capital dollars, during the first half of 2026, according to the article’s analysis of PitchBook data.

DailyOps Insight

Operational Impact: High — Anthropic’s IPO connects equity ownership, cloud commitments and accelerator diversity in one transaction. The story is not only about valuation; it is about who supplies the systems that frontier AI depends on.

Actionable Advice: When evaluating AI infrastructure, map the full dependency chain. Track investors, cloud providers, chip suppliers and workload-orchestration layers together. The most important risk or advantage may sit between those categories rather than inside any one vendor relationship.

Ops Takeaways

Sources