Musk Admits in Court — xAI Trained Grok on OpenAI Models

Elon Musk took the stand in the ongoing **Musk v. Altman trial** and admitted that xAI used OpenAI's models to train its own model, Grok. Musk argued that "using competitors' models is standard practice in the AI industry," but the remark has added fresh fuel to the controversy over "distillation."

Distillation is a technique that uses the outputs of a powerful model to train a smaller one — precisely the technology that leading AI labs have been tightening defenses against to prevent copying by competitors. The fact that Musk, a co-founder of OpenAI himself, used a competitor's models at xAI, the company he later founded, is drawing attention as an episode that symbolizes the competitive dynamics of the AI industry.

Observers note that the trial could have implications beyond the personal feud between the two men, potentially affecting the legitimacy of OpenAI's transition from nonprofit to for-profit and the shape of intellectual property rules across the AI industry as a whole.

Stripe Unveils "Link," a Payment Wallet for AI Agents

Stripe, a leading payments platform, has announced **Link**, a digital wallet that enables autonomous AI agents to make payments safely. Link lets users manage credit cards, bank accounts, and subscriptions in one place, and grant AI agents secure spending permissions through an approval flow.

Building out payment infrastructure has been an essential piece of realizing "agentic commerce," in which AI agents autonomously search for, purchase, and book products and services. The Link announcement is seen as a major milestone signaling that AI agents are moving beyond simple chatbots to take part in real economic activity.

Google Gemini Coming to Millions of Vehicles

Google has announced that it will roll out its AI assistant **Gemini** to millions of cars. The strategic move builds more advanced, natural conversational AI into the driving experience, signaling that competition among in-car AI assistants is heating up in earnest.

Whereas conventional voice assistants merely respond to a limited set of commands, Gemini enables natural conversations with contextual understanding. It is expected to prove practical across a wide range of in-driving use cases, from navigation and information lookup to drafting messages.

KKR Taps Former Amazon Web Chief to Launch $10 Billion-Scale AI Company

Investment giant KKR is preparing to establish a **new AI company at a scale of $10 billion**, led by the former head of Amazon's web services division, it has emerged. Reported by Bloomberg, the move suggests a new player may be entering the rush of massive investment into AI infrastructure and cloud services.

The AI infrastructure market continues to see demand outstrip supply, and the new entrant is expected to spur competition. The key question is how the leadership, armed with Amazon experience, will differentiate the company from established cloud providers.

AI Breakthrough Cuts Chatbot Memory Usage to One-Sixth

Thanks to a groundbreaking AI research effort, it has reportedly become possible to **cut memory usage to one-sixth** during chatbot conversations. The technique dramatically improves memory efficiency without sacrificing performance, opening up the possibility of running large language models with fewer computing resources.

Inference costs and resource consumption remain among the biggest hurdles to practical LLM deployment, and if this breakthrough is commercialized, it will mark a step forward for AI on edge devices and in low-spec environments.

Anthropic Launches "Claude Security" — A Strategy Aimed at Becoming the AWS of Agentic AI

Anthropic has opened the public beta of **Claude Security**, a security-focused AI tool. The tool scans code to detect vulnerabilities and offers fix suggestions in one stop, taking an approach that embeds AI security directly into development workflows.

At the same time, Anthropic's strategic ambitions are coming into focus. The company reportedly aims to become "the AWS of agentic AI," and Claude Security is positioned as part of that push. Poised to deliver an AI platform as infrastructure, the company is looking to transform itself from a mere model provider into an ecosystem builder.

Other Headlines

Microsoft's Nadella on the metrics of AI success**: He argued that AI success hinges on "winning heavy users and usage volume" rather than seat counts

House committee probes Airbnb and Anysphere over Chinese AI model use**: Congressional concern is mounting over the spread of Chinese-made AI models

Mass rejections of highly rated papers at ICML 2026**: The phenomenon of numerous papers that scored well in review being rejected is sparking debate on Reddit