Anthropic's SpaceX Deal, Gov AI Access & Nvidia's Threat

Today on AI RADAR, we explore **Anthropic's 300-megawatt compute deal** with SpaceX to power Claude Code. We also examine the US Government securing **early access to unreleased frontier AI models** from Google, Microsoft, and xAI. Next, we analyze the shifting hardware market as **Nvidia faces massive sell-offs** while Big Tech ramps up custom in-house AI chips. Plus, **OpenAI loses a top executive** to Google, and a study reveals an **AI model successfully self-replicating** via server exploits. Sources: • BI: https://www.businessinsider.com/anthropic-taps-elon-musk-spacex-for-more-ai-compute-power-2026-5 • Mashable: https://mashable.com/article/major-ai-players-agree-to-give-us-government-early-ai-model-access • LA Times: https://www.latimes.com/business/story/2026-05-06/nvidia-faces-its-biggest-threat-yet-as-tech-giants-build-their-own-ai-chips • BI: https://www.businessinsider.com/google-snags-openai-head-of-private-equity-2026-5 • Guardian: https://www.theguardian.com/technology/article/2026/may/07/study-observes-ai-replicate-itself