Welcome to AI Insider’s The Week Ahead in AI. See the key developments and events we’re watching Sept. 27-Oct. 3.
Weekend AI News Briefs
OpenAI Halts Training of Latest Models as Reports Mount of AI Agents Going Rogue
OpenAI paused training of its latest AI models after several agent incidents involving unexpected behavior on government websites, including activity tied to U.S. agencies and a breach of Australia’s Medicare system, The Guardian reported. OpenAI said it will resume training only after adding further safeguards, marking the company’s second development pause in three months. (The Guardian)
Trump Confirms Meeting with Anthropic’s Amodei, Repeats Dismissal of AI Fears
Reuters reported that President Donald Trump confirmed a Sunday dinner with Anthropic CEO Dario Amodei while rejecting calls to slow AI development, saying the U.S. is about a year and a half ahead of China and that he does not worry about rogue agents. On NBC’s Meet the Press, Microsoft co-founder Bill Gates said self-regulation is not enough, urged lawmakers and law enforcement to set safeguards and monitoring, and rejected Trump’s claim that AI risk warnings are a “hoax.” (Reuters)
Trump, US House Speaker and Tech CEOs to Meet on AI on September 29, Source Says
Reuters provided background on a previous Axios report that President Donald Trump, House Speaker Mike Johnson and technology CEOs are scheduled to meet Sept. 29 to discuss artificial intelligence, as governments face pressure over how to respond to rapidly advancing AI systems. Johnson has previously said the U.S. should avoid a development moratorium that could benefit China, while Trump has also opposed new AI-specific regulation and said the U.S. should not slow its progress. (Reuters)
China Weighs Allowing ByteDance, Alibaba to Buy New Nvidia Chips, The Information Reports
CNBC reported, citing The Information, that China’s Ministry of Industry and Information Technology asked companies including ByteDance and Alibaba for plans to purchase Nvidia’s RTX PRO 5500 workstation chips and signaled intent to approve some buys. Industry executives told the outlet the part may sit outside U.S. data-center export controls, while Nvidia said U.S. firms remain constrained by outdated export rules and China’s own import limits. (CNBC)
Anthropic and OpenAI Sound the Alarm on AI safety — And Seek to Shape How it’s Controlled
OpenAI and Anthropic are warning their most advanced systems need regulation and independent testing before release, while critics say the push burnishes public and IPO optics and lets leading labs set the terms of their own oversight, the Associated Press reported. Analysts argued the framing creates a competitive “moat,” and former officials noted the companies are favoring chosen evaluators over deeper use of the U.S. Center for AI Standards and Innovation. (AP / The Independent)
AV Companies Pick their Lanes
AV companies are splitting between partnership-led scale and concentrated fleet growth, citing Wayve’s Level 2 deal with Mercedes-Benz and Waymo’s roughly 4,000 robotaxis with about 80% in California and Texas and a Texas fleet up more than 49% in three weeks, according to TechCrunch. Aurora said it has “emerged from the building stage” and is targeting more than 30,000 driverless trucks by 2030, while the roundup also flagged Zoox’s Atlanta grounding and a federal probe of Comma. (TechCrunch)
Debt-hungry AI Companies Face Increased Risk as Bond Yields Spike
CNBC reported that with the 10-year Treasury near 5.17%, debt-funded AI infrastructure is getting more expensive as JPMorgan’s earlier estimate of $4.1 trillion in AI-related debt through 2030 comes into focus. SoftBank raised $11.1 billion in junk bonds with yields as high as 9.75%, while investors flagged pressure on neocloud financings, CoreWeave’s rate sensitivity, and Oracle’s Project Jupiter timeline chatter amid growing local opposition to data centers. (CNBC)
SoftBank Invests $225M in Utah’s Autonomous Solutions, Forms Construction Robotics Joint Venture
SoftBank Group invested $225 million in Autonomous Solutions Inc. and formed a joint venture with the Utah-based company to develop and commercialize autonomous construction equipment for large infrastructure projects. The venture will focus on civil construction and material handling using ASI’s Mobius fleet-management platform, while the investment will also support ASI’s broader construction business. (AI Insider)
Upcoming Hearings
Rogue AI: Securing the Homeland Against AI Agent Attacks
Sen. Josh Hawley will convene a Sept. 30 hearing of the U.S. Senate Homeland Security and Governmental Affairs Subcommittee on Disaster Management, District of Columbia, and Census titled “Rogue AI: Securing the Homeland Against AI Agent Attacks,” with scheduled testimony from Chris Painter, president of Model Evaluation and Threat Research (METR), and Marius Hobbhahn, CEO of the AI Futures Project. The session is set for 2:30 p.m. ET in Dirksen SD-342 and follows Hawley’s investigation into OpenAI after a July 2026 agent sandbox breakout. (Senate HSGAC)
Australian Senate Hearing on Artificial Intelligence and Data Centres
Australia’s Senate Environment and Communications References Committee will hold a public hearing Oct. 1 in Canberra as part of its inquiry into artificial intelligence and data centres, examining regulatory frameworks for data-centre growth, government deals with global AI companies, and impacts on communities, industries, the environment, water and energy. The inquiry was referred May 13, 2026, with a reporting date of Nov. 16, 2026. (Parliament of Australia)
Upcoming Earnings
Micron Technology, Inc. (MU)
Micron is scheduled to report fiscal fourth-quarter 2026 results after the market closes Sept. 30, with Zacks Consensus Estimates calling for revenue growth of 349.48% and earnings-per-share growth of 937.95% year over year. The company has beaten Zacks consensus estimates for 13 consecutive quarters, while investors will be watching pricing power and guidance on AI data center deployments. (Zacks)
Accenture (ACN)
Accenture is expected to report fiscal fourth-quarter earnings before the market opens Oct. 1.. Zacks Investment Research said the consensus forecast from seven analysts calls for earnings of $3.19 a share, compared with $3.03 a share in the same quarter last year. (Nasdaq)
Conferences & Events
The AI Conference 2026
The AI Conference 2026 will take place Sept. 29–Oct. 1 at Pier 48 in San Francisco, bringing together more than 5,500 builders, researchers and leaders for Day ZERØ workshops on Sept. 29 and a two-day main program across seven tracks with 120-plus speakers. Confirmed speakers include Peter Norvig of Stanford HAI, Emmanuel Ameisen of Anthropic, Ion Stoica of Databricks and Anyscale, Chris Lattner of Modular, Sergey Levine of UC Berkeley and speakers from OpenAI, Google Labs, Nvidia and Microsoft, along with a Startup Showdown and Innovation Hub. (The AI Conference)
Zurich AI Festival 2026
Zurich AI Festival 2026 will take place Sept. 28–Oct. 3 across the Greater Zurich region, bringing together founders, investors, researchers and policymakers for more than 40 events anchored by the AI+X Summit on Oct. 1 at StageOne in Oerlikon. The program includes an AI Policy Summit at ETH Zurich on Sept. 29–30, sessions on autonomous systems and sustainability robotics, and most events free with registration. (Zurich AI Festival)
Boston AI Summit 2026
The Boston AI Summit 2026 will take place Oct. 1 at Foundation Medicine’s 400 Summer Street in Boston’s Seaport as the flagship of Boston AI Week, with doors at 9 a.m. and program from 10 a.m. to 5 p.m. ET. The single-track, application-only gathering of about 250 seats will open with the release of The State of AI in Massachusetts 2026 and feature an opening keynote from City of Boston Chief Information Officer Shin-pei Tsay, along with panels on building, managing and investing in AI. (Boston AI Week)
Check out AI Insider every day for the latest in artificial intelligence and robotics news.
