Anthropic’s IPO to Benefit VCs Significantly, but Amazon to Gain Even More

Anthropic’s IPO: Major Gains Expected for Tech Giants

As the anticipated public offering for Anthropic approaches, numerous investors are poised for substantial returns. Venture capital firms such as Menlo Ventures, Sequoia, Sound Ventures, and Spark Capital are projected to reap significant profits, with some estimates suggesting they could see billions in returns.

Interestingly, according to Anthropic’s S-1 filing, none of these firms has reached the 5% ownership threshold typically necessary for inclusion in the principal stockholders’ table. This is unusual; in most tech IPOs, lead investors usually retain much larger stakes, often in double digits. The situation arises primarily due to Anthropic’s staggering $965 billion private valuation, where a 1% stake is equivalent to nearly $10 billion. For instance, while Menlo’s investment is reportedly valued at around $14 billion, it constitutes only about 1.5% of the company, as noted by recent figures from Bloomberg.

Although venture firms may lack substantial percentage holdings, Anthropic has secured significant investments from strategic partners, many of whom are also its business allies. Notably, Amazon emerges as potentially the largest benefactor from the IPO.

Amazon first invested in Anthropic in 2023, pledging up to $4 billion. Since then, the tech giant has continued to funnel more capital into the company, choosing to structure a large portion of its investment as convertible notes, which transform into shares upon an IPO or sale.

According to the S-1, Amazon currently holds convertible notes with an outstanding principal of $6.2 billion as of June 30. This amount constitutes Anthropic’s complete debt load. However, as Anthropic’s valuation soared through various funding rounds, the value of the shares those notes are set to convert into increased dramatically. As of June 30, Amazon valued its note holdings at approximately $97.9 billion.

With the rising value of these convertible notes, Anthropic had to reflect these adjustments on its balance sheet, recording a total of $72.4 billion for the first half of 2026 under “change in fair value of convertible notes, derivative liabilities, and SAFEs.” Importantly, this charge is non-cash, indicating that no actual funds were disbursed out of the company.

Amazon’s notes could convert into approximately 1.84 billion shares, according to the S-1. However, a limit exists, capping Amazon’s stake at 12% of Anthropic’s “partially diluted” share count—a number that factors in unissued shares such as stock options. Notably, Amazon has the option to waive this cap, allowing any non-converted notes to remain outstanding for future conversion.

In addition to these notes, Amazon has invested a total of $10 billion in non-voting preferred stock during the second quarter. This investment was achieved by injecting $5 billion into each of Anthropic’s last two funding rounds. When added to the $8 billion investment via notes from 2023 to 2025, Amazon’s combined contribution amounts to roughly $18 billion.

Apart from stock gains, Amazon is set to benefit from significant future revenue, as Anthropic has committed to spending at least $110 billion on AWS over the next decade.

Google is also positioned to gain considerably from both the IPO and ongoing cloud computing agreements with Anthropic. The technology giant held approximately 14% of Anthropic early last year, based on court records, and has invested additional capital since then. Analysts suggest that Google may currently hold around 5.6% of Anthropic, which has also secured at least $114 billion in cloud commitments from the AI startup through 2033.

While venture capitalists stand to make impressive returns from Anthropic’s IPO, the real winners may be tech giants like Amazon, who supported a potential competitor early on and are now its primary cloud service provider.

Importantly, the S-1 highlights potential conflicts of interest, cautioning that some cloud service providers also develop competing AI models, which could lead to misaligned objectives with Anthropic.