OpenAI and Oracle reportedly ink historic cloud computing deal
Roundup Sep 11, 2025
Hello data folks đ
Oracle and OpenAI signed a $300 billion compute contract spanning five years, with purchases starting in 2027, according to Wall Street Journal reporting. This would rank among the largest cloud deals in history.
The agreement follows OpenAI's strategic shift away from Microsoft Azure exclusivity, which began in summer 2024. OpenAI has also signed compute deals with Google despite competitive tensions, and committed $500 billion with SoftBank and Oracle for domestic data centers through the Stargate Project.
Source: Techcrunch
RSS co-creator launches new protocol for AI data licensing
Techcrunch | September 10, 2025 | 5 minute read
RSS co-creator Eckart Walther launched Real Simple Licensing (RSL), a new protocol for AI training data licensing at internet scale. The system enables machine-readable licensing agreements through robots.txt files and establishes collective licensing similar to ASCAP for musicians.
Major publishers including Reddit, Yahoo, Medium, and Quora have joined the initiative following Anthropic's $1.5B copyright settlement and 40+ pending AI copyright cases. The RSL Collective negotiates terms and collects royalties, though tracking usage remains technically challengingâespecially for inference-based payments versus blanket licensing fees.
Will Amazon S3 Vectors Kill Vector Databasesâor Save Them?
Zilliz | September 4, 2025 | 15 minute read
AWS launched S3 Vectors, offering vector storage at $0.06/GBâroughly 10x cheaper than traditional vector databases. The service handles up to 50M vectors per table with 500-700ms cold query latency and sub-200ms hot queries at 200 QPS (queries per second).
Testing reveals clear constraints: recall rates plateau at 85-90%, write performance caps under 2MB/s, and filtering can drop recall below 50%. The service works well for cold data archiving and low-QPS RAG applications but struggles with high-performance search and frequent updates.
PagerDuty quarterly financial report
PagerDuty investor relations | September 4, 2025 | 1.5 hour read
PagerDuty's Q2 filing highlights its platform strategy centered on AIOps and a new generative AI assistant, designed to ingest and correlate signals from over 700 direct integrations. The company positions its Operations Cloud as the core for unifying mission-critical operational data.
This focus on advanced data capabilities, however, is set against a 12% YoY decrease in R&D spending ($4.2M), attributed mainly to reduced engineering headcount in a push for profitability. The filing emphasizes using AI to help customers manage complexity âsmarter and fasterâ amidst this internal belt-tightening.
Two More Things
Thatâs the brief.


