第一财经

OpenAI's security model has been launched on the Amazon platform, and a $50 billion investment has been completed.

原文:OpenAI安全模型上线亚马逊平台,500亿美元投资已完成

Summary of Key Points

Amazon has been quite active in the AI field recently:

1. It collaborated with OpenAI to upgrade its capabilities, launching two new AI models for network security.

2. It made significant investments in both OpenAI ($50 billion) and Anthropic ($13 billion so far, with an additional potential investment of up to $20 billion), which not only resulted in substantial equity gains (over 60% of Amazon's second-quarter net profit came from Anthropic) but also contributed to the rapid growth of its AWS cloud business.

3. Amazon adopted a dual approach by investing in external AI companies while also developing its own models, reducing its reliance on any single technology.

Detailed Analysis

1. Amazon × OpenAI: Further Cooperation and New Network Security AI Models

The newly released Daybreak Red and Blue models are part of OpenAI's Daybreak initiative for network defense. Here’s a brief overview:

  • Daybreak Red: An AI model specifically designed to detect and respond to cyberattacks, utilizing GPT-5.6 Cyber technology.
  • Daybreak Blue: Builds on the foundation of general-purpose AI models like GPT-5.6 Sol by adding security features, making it suitable for defensive network security tasks.

This isn’t the first time Amazon has partnered with OpenAI. In February this year, Amazon invested $50 billion in OpenAI, and AWS became the exclusive third-party cloud provider for OpenAI’s cutting-edge models. Last July, models such as GPT-5.6 Sol were already available through AWS. This expansion into network security represents a win-win situation for both parties: OpenAI can reach more enterprise customers via AWS, while Amazon enhances its Bedrock platform (its AI model marketplace) and encourages more businesses to use its cloud services.

2. Profitable Investments in AI Companies

Amazon has made substantial investments in the AI sector:

  • OpenAI: The $50 billion investment was allocated over multiple phases—$150 billion in the first quarter and $137 billion in the second quarter, with the remaining $213 billion to be paid out later.
  • Anthropic: Amazon has already invested $13 billion (an additional $50 billion in April this year), with a potential total investment of up to $20 billion.

These investments have quickly paid off. Amazon’s second-quarter net profit soared by 244%, with $534 billion coming from the appreciation of its stake in Anthropic. In other words, Amazon’s initial investment in Anthropic has appreciated in value, contributing significantly to its profits. This is akin to “buying shares in an AI company and earning a substantial return without doing much effort.”

3. AI Investments Boosting AWS Growth

AWS is a crucial revenue source for Amazon, and these AI investments have further strengthened its position:

  • AWS’s sales in the second quarter reached $42.2 billion, a year-on-year increase of 37%, representing the fastest growth in 18 quarters.
  • There are a backlog of orders totaling $496 billion (a three-digit year-on-year increase), indicating substantial future business potential.
  • For example, Anthropic has committed to spending $100 billion over ten years on AWS technologies, which not only ensures long-term revenue for AWS but also promotes the use of Amazon’s self-developed chips (such as Trainium and Inferentia), helping to reduce costs.

Amazon plans to invest another $220 billion in AWS and generative AI by 2026, demonstrating its confidence in this area.

4. A Dual AI Strategy: Leveraging External Resources While Building Internal Capabilities

Amazon’s approach to AI involves a balanced strategy:

  • Leveraging External Resources: By investing in leading AI companies like OpenAI and Anthropic, Amazon gains access to cutting-edge technologies and ensures their use of AWS services.
  • Building Internal Capabilities: It also develops its own AI models, such as the Titan series on its Bedrock platform. CEO Andy Jassy emphasized that no single model will dominate the market forever. Developing in-house models allows Amazon to control costs (by avoiding licensing fees) and reduces reliance on any one company. This strategy helps mitigate risks and maintains a competitive advantage in the AI landscape.

In summary, Amazon has not only made profitable investments but also boosted its cloud business through AI. It is simultaneously strengthening its own AI capabilities, solidifying its position in the AI era. For users, this means a wider range of AI services on AWS and potentially lower costs for businesses utilizing AI technologies.