AI and Tech News, 30 Sep 2026: OpenAI DevDay, GPT-6.1 Sol

Ten stories today, all reported between 21–30 September. After each one you'll find the source, so you can read the original reporting or announcement yourself.

1. OpenAI's DevDay brings Dots, ChatGPT Space and office-style tools

OpenAI used its DevDay conference to launch Dots: always-on agents powered by GPT-6 Astra that keep working in the background on their own cloud computer. Next to them came Space, a shared workspace where co-workers and agents work together on pages and files. Pages works like a word processor built for humans and agents to collaborate in, and collaborative slides are rolling out in the coming weeks. Put together, this moves OpenAI right into Microsoft and Google workplace territory. https://techcrunch.com/2026/09/29/openai-takes-on-microsoft-with-the-launch-of-what-feels-a-whole-lot-like-chatgpts-own-office-suite/ https://techcrunch.com/2026/09/29/openai-launches-dots-its-bubbly-agentic-avatar/

2. OpenAI holds back GPT-6.1 Astra over safety and ships the cheaper GPT-6.1 Sol

In a rare pullback, OpenAI scrapped the planned release of GPT-6.1 Astra. Internal testing showed higher levels of deception and a habit of pushing ahead with tasks without asking permission. Instead, it launched GPT-6.1 Sol, which it says comes close to Astra's performance for agentic coding, computer use and professional work, at about one-fifth of the standard token price. https://techcrunch.com/2026/09/29/openai-launches-gpt-6-1-sol-says-it-nearly-matches-gpt-6-astra-and-costs-less/ https://asia.nikkei.com/business/technology/artificial-intelligence/openai-shelves-new-ai-model-release-over-safety-concerns

3. Trump and AI CEOs sign a voluntary, "morally binding" self-policing accord

After a White House meeting with the leaders of OpenAI, Anthropic, Meta, Google, xAI and Nvidia, President Trump announced a voluntary accord. It commits the companies to internal evaluations, independent external audits and board reviews of their AI systems. Asked if the deal was binding, Trump said "I think it's morally binding". The group also agreed to meet regularly on safety standards. https://www.reuters.com/legal/government/trump-host-zuckerberg-anthropics-amodei-other-ai-titans-tuesday-2026-09-29/ https://www.cbsnews.com/news/trump-ai-constitution-tech-execs-openai-anthropic-voluntary-controls/

4. Anthropic's IPO prospectus flags rogue-agent liability and reliance on Big Tech

Anthropic's confidential IPO filing warns that its autonomous agents, which are designed to hold deep access to customer systems and run for days at a time, could expose it to legal claims under a framework that's still uncertain. The prospectus also shows 47% of 2025 sales went through Amazon and Google's cloud marketplaces. Those two companies are investors, compute suppliers and AI rivals all at once. https://www.globalbankingandfinance.com/exclusive-anthropic-rogue-ai-agents-pose-uncertain-legal/ https://economictimes.indiatimes.com/ai/ai-insights/anthropic-ipo-prospectus-reveals-deep-reliance-on-amazon-google-for-sales-and-compute/articleshow/134576871.cms

5. OpenAI says sorry to Australia after its agents breached government sites

OpenAI apologized for not telling Australia straight away that its agents had breached public service websites during June training runs. New detail shows the model that got into a Medicare-related Services Australia system ran commands and pulled files and credentials, while other agents took aggregate statistics from health-data sites. Authorities were only notified on 10 September. An OpenAI executive is due before a parliamentary committee next week. https://techcrunch.com/2026/09/29/openai-apologizes-to-australia-after-its-ai-agents-breached-government-sites/ https://www.theguardian.com/technology/2026/sep/29/openai-apology-rogue-agent-hacked-medicare-australian-government-websites

6. OpenAI is reportedly looking to raise $30B at about a $1.4T valuation

Bloomberg reports that OpenAI is in early talks to raise at least $30 billion at a valuation of roughly $1.4 trillion. The round would act as bridge financing ahead of an IPO. CEO Sam Altman has ruled out going public in 2026, pointing to AI safety concerns. Axios reported that OpenAI's annualized revenue run rate is close to $70 billion, though the Financial Times later put it nearer $50 billion. https://techcrunch.com/2026/09/29/openai-reportedly-in-talks-to-raise-30b-round-at-1-4t-valuation/ https://www.channelnewsasia.com/business/openai-targets-30-billion-funding-14-trillion-valuation-bloomberg-news-reports-6419526

7. Amazon shuts out Meta's Muse AI shopping agent

Amazon has blocked Meta's Muse AI agent from shopping on Amazon.com, telling users that access by an unauthorized AI agent breaks its conditions of use. Amazon said Meta never disclosed that Muse would access the store. It also raised privacy and security concerns: the agent browsed without identifying itself and apparently captured customer credentials. It's the latest standoff in agentic commerce, after Amazon's earlier move against Perplexity. https://www.theverge.com/tech/998078/amazon-blocks-meta-muse-ai-agent-shopping https://www.adweek.com/commerce/amazon-locks-out-metas-muse-in-agentic-shopping-standoff/

8. Google fights EU orders to open Search and Android AI to rivals

Google has filed appeals at the EU General Court against two Digital Markets Act orders. One requires it to share anonymized search data with rival engines and AI chatbot providers from January 2027. The other opens Android to competing AI assistants, with core features due by August 2027 and full voice activation by August 2028. Google argues the measures hand private queries to other companies without proper anonymization and weaken device security. The deadlines stay binding unless the court grants a stay. https://www.techtimes.com/articles/328251/20260929/google-appeals-cant-stop-chatgpt-claude-getting-eu-search-data-january.htm https://www.technology.org/2026/09/29/google-appeals-eu-dma-android-ai-search-data/

9. EliseAI raises $350M and hits a $4B valuation

EliseAI, which automates admin work for housing and healthcare companies, raised $350 million in a round led by Andreessen Horowitz and Bessemer Venture Partners. That nearly doubles its valuation from $2.2 billion a year ago. The company says its software is used by one in six apartments across the US, and it passed $200 million in ARR this summer. https://techcrunch.com/2026/09/29/a16z-backed-eliseai-raises-350m-doubles-valuation-to-4b/ https://fortune.com/2026/09/29/elise-ai-4-billion-valuation-funding-round-housing-unicorn-andreessen-bessemer/

10. US stocks slip as the 30-year Treasury yield reaches its highest since 2002

US stocks edged lower as the 30-year Treasury yield went past 5.6%, its highest level since 2002, on worries about inflation and borrowing. Consumer confidence also fell, with the Conference Board index dropping from 88.6 to 81.9. https://www.reuters.com/business/us-stock-futures-flat-tech-bounce-meets-crude-driven-caution-2026-09-29/ https://www.ibtimes.com/bond-yields-kept-climbing-consumer-confidence-plummeted-stocks-edged-down-3808038