Topic: #nvidia
-
Google developing Frozen v2 chip for Gemini AI
According to reports, Google is developing a custom server chip codenamed Frozen v2 that bakes Gemini's architecture directly into hardware, projecting a 6–10x efficiency gain. The project remains exploratory with no official confirmation from Google, and deployment isn't targeted until 2028.
-
Nvidia Vera Rubin AI platform on schedule, customer testing underway
Nvidia's next-generation Vera Rubin AI platform is on track for production ramp-up in May 2026, with customer sampling already underway at AWS, Google Cloud, and Microsoft. The platform promises up to 10x reduction in inference costs versus Blackwell.
-
Nvidia shifts to revenue-sharing model with Australian cloud provider SharonAI
Nvidia announced a six-year partnership with Australian cloud provider SharonAI under a new revenue-sharing model, marking a shift toward predictable, subscription-like income streams from AI infrastructure deployments.
-
Nvidia Opens South Korea R&D Center, Strengthens Samsung and SK Hynix Ties
Nvidia announced hiring for a new research and development center in South Korea during CEO Jensen Huang's June 2026 visit, focusing on physical AI, robotics, and infrastructure solutions alongside discussions with the government and Hyundai Motor Group.
-
Nvidia CEO Huang: AI tokens now profitable, signaling shift to inference-driven revenue
Jensen Huang declared during Nvidia's Q1 FY2027 earnings call that tokens—units of AI model output—are now profitable for companies, marking a fundamental shift where inference has become a revenue engine rather than just a cost of business.