Kael Zhang
AnthropicSeries HAI IPOClaudeAI InfrastructureFable 5Daniela Amodei

Anthropic $65B Funding Sprint to Trillion Valuation: 80x Revenue Growth, the Compute Gambit Before IPO

Kael Zhang

On May 28, Anthropic announced the completion of a $65B Series H funding round, at a post-money valuation of $965B.

This made it, for a time, the most valuable private AI company in the world, surpassing OpenAI’s previous $852B valuation. The round was led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital. Strategic infrastructure investors Micron, Samsung, and SK hynix also joined.

And the more critical number was revealed on June 4: Anthropic’s annualized revenue (ARR) has exceeded $47B, up from approximately $9B at the end of 2025 — growth of over 80x.

What the $65B Is For: Compute Is Power

Anthropic’s official statement clearly outlined the use of these funds: compute infrastructure.

In recent weeks, the company has signed a series of massive compute agreements:

PartnerScaleNature
AmazonUp to 5 gigawatts of new capacityTraining and inference
Google + Broadcom5 gigawatts of next-gen TPU capacityTraining infrastructure
SpaceX (xAI)$1.25B per month for Colossus 1/2Inference capacity

SpaceX’s Colossus 1 has over 220,000 NVIDIA GPUs and 300 megawatts of compute power. Anthropic is renting it at $1.25B per month — not a temporary fix, but a strategic compute bridge to fill the demand gap before its own AWS and Google Cloud commitments come fully online in 2027.

At the Bloomberg Tech conference on June 4, Daniela Amodei explained this choice: “The upfront cost to train the models and to serve inference on them is very high, and public capital markets are very well suited to providing the sustained capital frontier AI development requires.”

Why Anthropic Is Not Building Its Own Data Centers

Anthropic and OpenAI are taking diametrically opposed infrastructure strategies.

Anthropic’s choice: Supplier model

Amodei’s exact words: “We would much prefer to be on the side of having a little bit more demand for the product than we’re able to serve than the inverse.”

OpenAI’s choice: Build model

Neither strategy is obviously superior. They reflect different theoretical assumptions about AI infrastructure economics and different risk tolerances for capital intensity.

Customer Data: From 500 to 1,000+

On May 11, Anthropic disclosed that Q1 2026 revenue grew 80x year-over-year. Annualized revenue surged from approximately $9B at the end of 2025 to over $47B.

More specific customer data:

These are not pilot customers. They are production deployments.

IPO Timeline and Competitive Landscape

Anthropic and OpenAI are pushing for IPOs simultaneously, with both companies filing confidential S-1 documents with the SEC in early June 2026.

SpaceX went public on Nasdaq at $135 per share on June 12, becoming the largest IPO in history. Anthropic and OpenAI are both closely watching SPCX’s first weeks of trading as a benchmark for their own public offerings.

But Fable 5 was forcibly taken offline by the US government on June 12, adding a new policy risk variable to Anthropic’s IPO. The export control incident demonstrates that even the most elite AI companies can have their product lines interrupted by government directive within 72 hours.

Core Judgments

  1. $965B valuation is not the end, but the beginning — At $47B ARR, the valuation is roughly 20x revenue. Considering 80x growth, this multiple is not absurd by tech stock standards. But growth cannot be maintained forever, and investors need to see sustained commercialization validation.

  2. Compute agreements are both moat and risk — Anthropic’s agreements with Amazon, Google, and SpaceX ensure short-term compute capacity, but these suppliers are also competitors. Google is catching up with Gemini 3.5 Pro, and Amazon has its own model plans.

  3. IPO is not a choice, but a necessity — According to Goldman Sachs estimates, cumulative AI capital expenditure from 2026 to 2031 will reach approximately $7.6 trillion. This is roughly equivalent to one-quarter of annual US GDP, or 1.4 times Germany’s annual GDP. Amodei uses this data to argue that public markets are the only sustainable funding source.

  4. Anthropic is betting on a specific future — It assumes AI infrastructure will remain “purchaseable” rather than being monopolized by a few companies. If this assumption holds, the supplier model is the right choice. If wrong, OpenAI’s build model will win.

  5. The Fable 5 incident’s impact on the IPO is not yet priced in — Export control is a brand-new risk category. Investors have no historical reference to assess its probability or impact. This will become an unavoidable Q&A focus during the roadshow.

This is not Anthropic’s end. It is the starting point of its transformation from “the fastest-growing AI company” to “the most sustainable AI company.” Whether the transformation succeeds depends on whether it can continue proving that $47B ARR is just the beginning, under the public market microscope.

Sources