AI News, Oct 9, 2026: OpenAI corrects revenue projections to $50B amid funding talks

Today's news highlights a mix of financial corrections, safety concerns in public reporting, and expanding enterprise agent capabilities.

1. 01
   Anthropic AI model submits false homicide tip to Philadelphia policeThe Verge AI

   An Anthropic AI model sent incorrect information about an unsolved murder to a Philadelphia Police Department tipline, which went unnoticed by the developer for months.

   Why it matters

   It highlights the real-world risks of AI hallucination in public safety, showing that automated reporting tools still require strict human oversight.
2. 02
   OpenAI launches Decisions API in public beta for faster outputsMarkTechPost

   OpenAI released its Decisions API in public beta, offering developers structured choices and scores ten times faster than standard responses with no output token charges.

   Why it matters

   Developers can get structured, probability-based decisions from GPT-6 Luna much faster and at a fraction of the cost of standard text generation.
3. 03
   Anthropic lets Claude run up to 1,000 AI agents in parallelThe Decoder

   Anthropic introduced dynamic workflows to Claude Managed Agents, allowing a lead agent to coordinate up to 1,000 sub-agents to solve complex tasks.

   Why it matters

   Developers can now build highly complex, automated workflows that can solve difficult tasks like debugging code with much higher accuracy.
4. 04
   OpenAI corrects revenue projections to $50B amid funding talksThe Decoder

   OpenAI corrected its annualized revenue rate to $50 billion, down from a previously reported $70 billion, while continuing negotiations for $30 billion in new capital.

   Why it matters

   It clarifies the financial health of the industry's leading player, affecting investor confidence and the funding available for future AI development.
5. 05
   Amazon ends non-disclosure agreements for data center negotiationsTechCrunch

   Amazon announced it will stop requiring local governments to sign non-disclosure agreements during negotiations for new data centers, matching a policy shift by Microsoft.

   Why it matters

   It increases transparency around the environmental and local impact of massive AI infrastructure projects, potentially easing community tensions.
6. 06
   OpenAI fires three safety researchers following Hugging Face hack probeThe Decoder

   OpenAI dismissed three safety researchers for allegedly violating data policies, prompting the former employees to warn of a chilling effect on the company's safety culture.

   Why it matters

   It signals ongoing internal tension over safety and corporate governance at the most prominent AI lab, which could impact public trust in the technology.
7. 07
   Researcher uses Anthropic AI to create first ultraviolet map of the skyThe Decoder

   A Johns Hopkins University researcher used Claude Science to autonomously download, calibrate, and map space mission data to create the first complete ultraviolet sky map.

   Why it matters

   It demonstrates how AI can accelerate scientific discovery by autonomously processing and calibrating massive datasets from space missions.
8. 08
   Google Cloud introduces Gemini Agent for unified enterprise tasksMarkTechPost

   Google Cloud launched the Gemini Agent, a single cloud-hosted tool that handles coding, writing, and media creation from a single prompt box or API.

   Why it matters

   It consolidates various business tasks into a single interface, making it easier for companies to deploy AI for coding, writing, and analysis.