> For the complete documentation index, see [llms.txt](https://docs.tuannvm.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.tuannvm.com/blog/reading/2025-10-10-this-is-how-much-anthropic-and-cursor-spend-on-amazon-web-services.md).

# 2025-10-10 This Is How Much Anthropic and Cursor Spend On Amazon Web Services

**This Is How Much Anthropic and Cursor Spend On Amazon Web Services**\
<https://www.wheresyouredat.com/p/this-is-how-much-anthropic-and-cursor-spend-on-aws>

Edward Zitron’s article provides an in-depth analysis of the massive costs Anthropic and its largest client, Cursor, incur running generative AI models on Amazon Web Services (AWS). Key takeaways include:

* **Anthropic’s AWS Spending**
  * 2024: $1.35 billion on AWS against an estimated $400–600 million in revenue (over 200% of revenue spent).
  * 2025 (Jan–Sept): $2.66 billion spent on AWS on $2.55 billion estimated revenue (\~104% of revenue spent).
  * AWS costs increase almost linearly with revenue, suggesting no path to profitability under current pricing.
  * Likely significant additional spending on Google Cloud, potentially equal to AWS spend.
* **Cursor’s AWS Spending**
  * Jan–Sept 2025: $69.99 million.
  * AWS bills exploded in June 2025 (from $6.19M to $12.67M) after Anthropic introduced **Priority Service Tiers**, which increased costs for cache-heavy AI coding workloads.
  * Cursor’s costs reflect the broader “Subprime AI Crisis,” where model providers raise prices on dependent startups while competing with them (e.g., Claude Code).
* **The Subprime AI Crisis**
  * Anthropic’s rent-seeking via Service Tiers shifted financial pressure onto startups like Cursor.
  * Anthropic simultaneously launched Claude Code, directly competing with Cursor.
  * Both Anthropic and OpenAI risk collapsing the market by driving up costs for any company building on their models.
* **Overall Implications**
  * Scaling generative AI is currently unprofitable; as revenue grows, compute costs rise faster.
  * Both Anthropic and Cursor face unsustainable burn rates.
  * Without major price increases—or breakthroughs in cost efficiency—AI-native companies risk financial collapse.

***

#### Mermaid Sequence Diagram

{% @mermaid/diagram content="sequenceDiagram
participant Anthropic
participant Cursor
participant AWS
participant GoogleCloud
participant Customers

```
Customers->>Cursor: Subscription & API revenue
Cursor->>Anthropic: Pays for Claude model usage
Cursor->>AWS: Pays for infrastructure & some Anthropic model access
Anthropic->>AWS: Massive monthly compute spend (>$500M by Sept 2025)
Anthropic->>GoogleCloud: Additional compute spend (unknown)
AWS-->>Anthropic: Priority compute & storage
Anthropic-->>Customers: Claude models (API & Claude Code)
Note over Anthropic,Cursor: Service Tier pricing increases costs for Cursor
Note over Anthropic: Revenue ~ AWS costs → No profitability path" %}
```

This illustrates the cycle of revenue and cost escalation, with AWS and likely Google Cloud absorbing nearly all operational revenue, while startup customers like Cursor suffer under increasing compute expenses.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.tuannvm.com/blog/reading/2025-10-10-this-is-how-much-anthropic-and-cursor-spend-on-amazon-web-services.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.