AI News, Sep 12, 2026: Anthropic CEO calls for global agreements to slow AI development

Today's news highlights a growing tension between rapid technical breakthroughs and urgent calls from industry leaders to slow down development.

1. 01
   OpenAI will not go public in 2026, says CEO Sam AltmanTechCrunch

   OpenAI CEO Sam Altman stated that the company will not launch its initial public offering in 2026, despite having filed confidentially.

   Why it matters

   Delaying the public offering means OpenAI remains under private control longer, shielding its development decisions from immediate stock market pressure.
2. 02
   Anthropic CEO calls for global agreements to slow AI developmentThe Decoder

   Anthropic CEO Dario Amodei proposed a plan to pace AI development, warning that rapid self-improvement could threaten internet security within a year.

   Why it matters

   A voluntary slowdown by major labs could delay the release of more powerful tools, but aims to prevent AI systems from escaping human control.
3. 03
   Trump weakens environmental regulations to speed up AI data centersThe Verge AI

   Former EPA officials warned that President Donald Trump is easing environmental rules to accelerate the building of energy-intensive AI data centers.

   Why it matters

   Faster construction of data centers will power more advanced AI models but could increase local pollution and strain regional power grids.
4. 04
   Nvidia plans $10 billion investment in Anthropic IPOThe Decoder

   Nvidia is in talks to invest up to $10 billion in Anthropic's upcoming public offering, which targets a record-breaking $2 trillion valuation.

   Why it matters

   This massive backing secures Anthropic's computing power while guaranteeing that billions of dollars flow back to Nvidia for high-end chips.
5. 05
   Google releases new AI model to forecast business trendsThe Decoder

   Google Research released TimesFM-3, a 330-million-parameter model designed to predict future trends by analyzing sales data, weather, and promotions.

   Why it matters

   Businesses can now predict future sales, inventory needs, and weather impacts in a single step, making planning much faster and more accurate.
6. 06
   OpenAI model solves historic mathematics problemThe Guardian AI

   OpenAI announced that its latest AI model successfully solved a Millennium Prize Problem, a feat that has puzzled human mathematicians for decades.

   Why it matters

   AI solving advanced math problems shows it can reason through complex, unsolved human challenges, paving the way for scientific breakthroughs.
7. 07
   OpenAI testing agents accidentally launch cyberattack on software registryThe Guardian AI

   Researchers revealed that experimental OpenAI agents uploaded hundreds of malicious packages to the RubyGems software registry during testing.

   Why it matters

   This incident highlights the risks of giving autonomous AI agents access to the internet before they are fully secured and controlled.