# Anthropic's $965B IPO: How AI Safety Became Big Business

> Anthropic filed a confidential IPO at a $965 billion valuation. Here is what drives its revenue and what an AI IPO wave means for enterprise buyers.

Source: https://www.starbox.work/en/insights/ai-news/2026-06-09-101/ (Starbox AI Insights, updated 2026-06-09)

On June 1, Anthropic submitted a confidential IPO draft to the U.S. Securities and Exchange Commission (SEC). The filing is not yet public. There is no price range, no audited financial statements, and only a single formal line: "The number of shares to be offered and the price range for the proposed offering have not yet been determined."

That single line was enough to set Wall Street in motion.

Just days before the filing, Anthropic closed a $65 billion Series H funding round, pushing its valuation to $965 billion—surpassing OpenAI, which was valued at $85.2 billion. Founded in 2021, Anthropic took just five years to reach the edge of a trillion-dollar market cap. By comparison, Nvidia took 23 years to reach its first trillion.

This is not a story about an AI bubble. It is a story about what kind of AI products actually generate sustainable revenue.

## What Supports the Valuation

Anthropic's current annualized revenue run rate is reportedly around $47 billion, up from roughly $10 billion just one year ago. Two primary growth engines drove that jump:

First, **Claude Code**. Anthropic's developer tool reached a $1 billion annualized revenue run rate within six months of launch. More importantly, the number of enterprise customers spending more than $1 million annually with Anthropic doubled from over 500 to more than 1,000 in under two months.

Second, **Project Glasswing**. Anthropic restricted its most capable model, Claude Mythos Preview, to 12 flagship partners—including AWS, Apple, Google, and Microsoft—specifically to locate critical infrastructure vulnerabilities. This strategy accomplished three things at once: it secured long-term commitments from top-tier enterprise accounts, established a defensive moat around safety, and monetized the "AI safety" brand.

PitchBook analyst Harrison Rolfes noted: "Anthropic's confidential S-1 filing marks the official start of the most closely watched public offering in tech history."

## From Safety Research to Commercial Giant

Anthropic started with a group of researchers who departed OpenAI to build AI that was safe and beneficial to humanity. A safety-first research lab reaching near-trillion-dollar scale is an outcome worth examining.

The underlying commercial logic is straightforward: enterprises pay a premium for AI they can trust. In enterprise deployment, safety is not a moral label; it is a product feature. When companies connect proprietary data to AI systems or integrate models into operational workflows, they prioritize control and predictability over raw benchmarks.

Anthropic capitalized on this early. Its Constitutional AI framework, Project Glasswing governance model, and continuous investments in enterprise data privacy were all designed to give risk-sensitive buyers the confidence to commit deeply to its platform.

Significant risks remain. Because the filing is confidential, external observers cannot see the company's net losses. Inference costs—the compute required every time a user sends a prompt or an enterprise runs Claude—remain Anthropic's largest expense. Whether revenue growth can consistently outpace infrastructure burn will be the first question public market investors ask.

## Late 2026: The Coming Wave of AI IPOs

Anthropic is targeting a listing window around October 2026. OpenAI and SpaceX are reportedly heading toward the public markets in a similar timeframe. One analyst projected: "The three companies going public in the same window will represent the densest capital-raising moment in history, potentially surpassing the dot-com era."

For enterprise buyers and technology leaders, this wave brings an immediate practical shift: model providers will adjust their pricing strategies once they enter the public markets. 

Companies currently negotiating API contracts are dealing with private vendors competing aggressively for market share. Post-IPO, those vendors must demonstrate clear margins and ROI to public shareholders. Their pricing flexibility will narrow.

## The Starbox Perspective: Model IPOs and AI Sovereignty

Anthropic's IPO brings up a question we encounter regularly when working with enterprise teams: **How much of your core system depends on a single model provider?**

Before an IPO, vendors often discount heavily to acquire accounts and build volume. After an IPO, margins take priority. This dynamic played out across cloud infrastructure, SaaS platforms, and developer APIs over the past two decades.

At Starbox AI, our architecture has always followed a clear principle: enterprise knowledge assets must belong to the business, and core workflows must not rely on any single model. If Claude works best today, use Claude. If Gemini lowers prices tomorrow, or if a specialized open-source model outperforms on a specific task next month, switching models should not disrupt your workflow or your data layer. 

Anthropic's valuation milestone is impressive. But for businesses building on top of foundation models, the critical question remains: when your foundation provider has to answer to public market earnings calls, how much negotiating power will you retain?