ATLAS News · Technology
Bloomberg: Amazon and Google to hold about 28% of Anthropic after IPO
Bloomberg reports Amazon and Google would own about 28% of Anthropic after an IPO, while ETFs, swaps and perps offer indirect pre-IPO exposure.
What happened
Bloomberg reports that Amazon and Google are expected to hold about 28% of Anthropic's stock once the company completes an initial public offering.
In a separate story, Bloomberg says investors can already gain indirect exposure to Anthropic before any listing. It names exchange-traded funds, swaps and perpetual futures, known as perps, as routes to what it calls a back door to Anthropic stock.
The sources provided do not give further detail on the timing of the IPO or how each instrument works.
Context
Anthropic is an artificial intelligence company, and both Amazon and Google are among its known backers.
Why it matters
The reports show how ownership could be split among large technology investors at listing and how some market participants seek exposure to a company before it trades publicly.
Sources
ATLAS writes its own summaries from the reporting of these outlets. Each link opens the original publisher.
- Amazon, Google to Have About 28% of Anthropic Stock After IPO · Bloomberg.com · November 17, 2026
- ETFs, Swaps, Perps Offer Back Door to Anthropic ‘Stock’ Pre-IPO · Bloomberg.com · November 17, 2026
Related
- US House panel seeks Microsoft briefing on reported ByteDance AI deal talks
- Nvidia's Kress Dismisses Customer-Shift Talk as Broadcom, AMD Deal Share Is Cited
- ServiceNow Shares Slide as AI-Driven Software Selloff Spreads
- Microsoft Outlook Hit by Thousands of Problem Reports on Wednesday
- Microsoft Faces Canada Lawsuit and Nadella Apology Over Outage
What you get with ATLAS
Create my ATLAS account →Fair value, probabilities and earnings for every stock. Free account, no card. Information, not advice.
Information, not advice. This content is informational and is not an investment recommendation or an offer to buy or sell securities. ATLAS estimates come from statistical models using public data, can be wrong, and past performance does not guarantee future results. Disclaimer