September 17, 2026
Today’s AI news involves driverless taxis, an overloaded chip pipeline, and King Charles reminding the tech industry that humanity should probably survive the next product launch. Here are five developments worth your attention, with links to the reporting behind them. For more headlines throughout the day, visit my site The Drone Report, where AI news updates every half hour. Because apparently the future no longer takes a lunch break.
1. Lucid and Bolt plan 25,000 robotaxis for Europe
Lucid and ride-hailing company Bolt announced plans to deploy 25,000 autonomous vehicles across European cities. The partnership targets driverless operation within defined conditions and areas, with regulators and technology partners involved in the rollout. These are planned vehicles, not 25,000 taxis suddenly arriving Thursday afternoon. My take: the real test is handling ordinary transportation reliably, including roadworks and passengers who enter the wrong destination. Knight Rider made this look considerably easier, although KITT rarely had to find someone outside an airport Starbucks. Read more at Reuters.
2. Huawei says it cannot make enough AI equipment
Huawei says demand for its AI computing equipment in China exceeds its production capacity, limiting overseas expansion. The company also announced earlier release targets for two next-generation chips in 2027. These are company statements and plans, rather than independent proof that it has overtaken Nvidia. My take: the AI race increasingly depends on who can manufacture and deliver the hardware. You can promise the world’s smartest assistant, but somewhere a purchasing manager still has to ask when the parts arrive. Even the future gets backordered. Read more at Reuters.
3. Amazon joins the safety debate without endorsing a slowdown
Amazon says AI models should undergo rigorous testing and have strong safeguards before release. However, it stopped short of supporting an industry-wide development slowdown. Its comments, reported Thursday, add another position to the debate over how companies should manage increasingly capable systems. My take: testing before shipping sounds sensible. The harder question is what counts as passing and who gets to decide. “Customers who deployed this model also purchased an emergency shutdown switch” should remain a joke, not a product recommendation. Read more at Reuters.
4. A federal website’s AI search creates an awkward moment
Reuters reports that the Federal Register website used Alibaba’s Qwen model for a search feature removed Wednesday. The discovery follows FBI allegations that Alibaba copied Anthropic’s technology. Those remain allegations. Experts cited by Reuters did not identify an immediate security threat; the material was public, and risk depends partly on how processing was handled. My take: this is a procurement and policy consistency story before it is a spy thriller. Washington’s AI messaging and its software choices apparently need to be introduced at the next staff meeting. Read more at Reuters.
5. King Charles delivers a royal warning to AI leaders
King Charles told executives from companies including Nvidia, Google DeepMind, OpenAI, and Anthropic that AI needs safeguards against potentially catastrophic misuse. Speaking Thursday, he urged them to keep the technology serving humanity, communities, and the natural world. This was an appeal, not a new regulatory measure. My take: when the monarchy joins the AI safety discussion, the conversation has officially escaped Silicon Valley. Somewhere between The Crown and The Terminator, we have reached a scene where the king asks the engineers whether they checked the off switch. Read more at Reuters.