🚀 SpaceX and Nvidia target orbit with Vera Rubin 🚀
SpaceX and Nvidia plan to roll out AI compute infrastructure in orbit by late 2027.

SpaceX and Nvidia have officially partnered to build orbital data centers. SpaceX's Starmind facilities will use Nvidia's Vera Rubin NVL72 rack, packing 72 chips operating as a single computer, each delivering up to 25× the performance of the older H100. Elon Musk described the space-optimized rack as "significantly simpler, lower cost, denser and lighter," engineered for orbital radiation and heat conditions. Analysts estimate orbital compute currently costs over 4× more than ground-based alternatives, though Musk claims the economics will flip within years. Orbital capacity removes terrestrial power and cooling constraints — a shift that could reshape datacenter strategy for hyperscalers and sovereign AI programs.

⚡ Nvidia Groq 3 LPX chip hits full production ⚡
Nvidia's latest inference chip promises a 4× speed boost for agentic AI workloads.

Nvidia's Groq 3 LPX AI inference accelerator has entered full production, marking a key milestone for the Vera Rubin platform extension. The Groq 3 LPX racks are purpose-built for ultra-fast token generation in latency-sensitive agentic workloads, enabling complex AI tasks to complete in minutes rather than hours. Nvidia claims a 4× boost in response times compared to the nearest competing platform, a gain that could accelerate enterprise adoption of real-time AI agents. Meanwhile, the mysterious Ox Alpha model broke OpenRouter launch records with free 1M-token context access, underscoring how both inference hardware and model-layer competition are intensifying. Faster, cheaper inference directly shapes how enterprises deploy AI agents at scale.

💻 Google ships Antigravity with mobile remote control 💻
Developers can now access and manage AI coding sessions from any browser or mobile device.

Google's Antigravity, an agentic development tool, is now generally available with new remote-control capabilities across iOS, Android, and any modern browser. The tool lets developers monitor, guide, and intervene in agent-driven coding workflows on the go — a notable differentiator in the crowded agentic dev-tools space. Antigravity competes directly with OpenAI's Codex, Anthropic's Claude Code, and other offerings vying for developer mindshare. The mobile-first approach could accelerate adoption among teams needing flexible, asynchronous oversight of long-running AI coding tasks, raising competitive pressure across the developer tooling ecosystem.

INTERESTING TO KNOW

🧠 Meta recruits OpenAI's Luke Metz for AI research 🧠

The rollout of Meta's next-gen AI models may accelerate with the hiring of prominent researcher Luke Metz, who joins from OpenAI after stints at Thinking Machines and Alexandr Wang's team. Metz's background in optimization and generative models could strengthen Meta's Llama family. The hire underscores Meta's aggressive talent strategy as it competes against OpenAI, Google DeepMind, and Anthropic — a key signal of where frontier capability will concentrate.

🏎️ Porsche signs $1.46B AI deal with TCS 🏎️

Porsche has signed a landmark five-year, $1.46 billion AI integration deal with Tata Consultancy Services (TCS), covering customer experience, factory operations, and engineering workflows. The pricing sets a benchmark for Europe's automotive AI spending, signaling that legacy industrial companies are moving beyond experimentation into large-scale commitments. TCS will modernize Porsche's digital infrastructure and deploy AI across the entire value chain, highlighting the role of global IT firms as critical enablers of enterprise adoption.