AI 日报

AI 日报 · 2026-08-18

这期日报从 61 条资讯中筛选出 4 条重点 AI 新闻。关注主题集中在 openai、ai-infrastructure、ai-training-data。如果只先读两条，可以从《OpenAI signs record Ohio data center lease with Nvidia backing up to $105 billion》、《Stripe is reportedly acquiring AI startup OpenRouter for more than $7 billion》开始。

Overview

当天导读

从 61 条资讯中筛选出 4 条

OpenAI signs record Ohio data center lease with Nvidia backing up to $105 billion

This is a major industry announcement involving a record-scale AI data center lease, a massive

关联报道 1

Stripe is reportedly acquiring AI startup OpenRouter for more than $7 billion

A potentially major AI infrastructure acquisition that could significantly affect model access,

关联报道 1

AirTag reveals how Amazon destroys rare books for AI training

This is a noteworthy AI/data-rights controversy involving alleged destructive book scanning for

关联报道 1

Anthropic explains how Claude’s invisible text watermarks will work

This is a high-value product and policy update: Anthropic is adopting SynthID-Text-style

关联报道 1

Stories

当日精选 8 条

The Decoder

8月17日 22:13 UTC·#openai

OpenAI signs record Ohio data center lease with Nvidia backing up to $105 billion

OpenAI has signed a record-breaking 20-year Ohio data center lease for about 8 GW of capacity, with Nvidia backing up to $105 billion and becoming the exclusive chip supplier.

This is a major industry announcement involving a record-scale AI data center lease, a massive Nvidia-backed financing commitment, and signals about infrastructure bottlenecks and off-balance-sheet AI spending. It has high potential impact on AI infrastructure, capital allocation, and chip supply, though the piece is largely a news report rather than a technical deep dive.

OpenAI signs record Ohio data center lease with Nvidia backing up to $105 billion Key Points - OpenAI is leasing a new data center with around 8 gigawatts of IT capacity from SoftBank subsidiary SB Energy in Ohio for 20 years. - Nvidia is backing the project with up to $105 billion and becomes the exclusive chip supplier. Nvidia CEO Jensen Huang now sees land, power, and building shells as the biggest bottlenecks in the AI buildout. - The Wall Street Journal reports that major tech companies hold around $3 trillion in AI commitments off their balance sheets.

查看单篇正文查看原文

The Decoder

8月17日 14:49 UTC·#ai-infrastructure

Stripe is reportedly acquiring AI startup OpenRouter for more than $7 billion

Stripe is reportedly acquiring multi-model AI gateway OpenRouter for more than $7 billion, shortly after its $1.3 billion Series B valuation.

A potentially major AI infrastructure acquisition that could significantly affect model access, routing, pricing, and competition, especially given OpenRouter's reported eight million users and support for more than 400 models; however, the claim is unconfirmed and no discussion or comment-quality data is provided.

Stripe is reportedly acquiring AI startup OpenRouter for more than $7 billion Stripe is reportedly acquiring the AI startup OpenRouter for more than $7 billion, according to Bloomberg. OpenRouter helps customers pick different AI models for various tasks based on their needs and budget, and the startup had just closed a Series B funding round of $113 million in May at a valuation of $1.3 billion. Investors include Sequoia, Andreessen Horowitz, Menlo Ventures, and Alphabet's Capital G.

查看单篇正文查看原文

The Decoder

8月17日 22:43 UTC·#ai-training-data

AirTag reveals how Amazon destroys rare books for AI training

An investigation reports that Amazon is allegedly buying printed rare books, scanning them for AI training data, and destroying the physical copies in the process.

This is a noteworthy AI/data-rights controversy involving alleged destructive book scanning for model training, with potential implications for copyright, archival preservation, and training-data sourcing. The piece is newsy rather than deeply technical, but it touches an important industry practice; no comments were provided to assess discussion quality.

AirTag reveals how Amazon destroys rare books for AI training Amazon buys printed books in bulk, scans them for AI training data, and destroys them in the process. That's what an investigation by 404 Media found after placing an AirTag inside a shipment of rare books and tracking it across the country. The shipment ended up at an Amazon warehouse in Las Vegas, home to a team called VGT3 whose logo features a Tyrannosaurus rex holding a book. Workers there say they cut off book spines to speed up scanning, destroying the copies in the process. Amazon uses the scanned data to train its Nova models.

查看单篇正文查看原文

The Verge AI

8月17日 18:57 UTC·#ai-compliance

Anthropic explains how Claude’s invisible text watermarks will work

Anthropic says Claude will use invisible SynthID-Text-style watermarks for generated text to satisfy EU AI Act transparency rules without affecting output quality or user cost.

This is a high-value product and policy update: Anthropic is adopting SynthID-Text-style invisible watermarks for Claude to align with EU AI Act transparency requirements. It’s relevant to AI tooling, provenance, and compliance, though it appears to be an implementation announcement rather than a major research breakthrough. No comments/discussion were provided to assess community reaction.

Anthropic has clarified how it’s planning to apply invisible watermarks to Claude-generated text in order to comply with Europe’s AI transparency rules. On Friday, Anthropic announced that Claude’s text marking system is “a version of the SynthID-Text approach” — an open-source watermarking technology developed by Google DeepMind that creates detectable patterns using wording probabilities. This watermarking feature, alongside C2PA support for Claude-processed images, is being introduced to meet Anthropic’s obligations under the European Union’s AI Act, which requires synthetic audio, image, video, and text to…

查看单篇正文查看原文