Nvidia unveiled its 550 billion-parameter open AI model Nemotron 3 Ultra and said it has started mass production of its next AI server platform Vera Rubin. CEO Jensen Huang introduced both at a GTC Taipei 2026 keynote. Nvidia said the model will be released in open form this week and highlighted competitiveness in benchmarks and token generation speed, while an external index score trailed a leading Chinese model.

Squeezebits, a startup focused on AI model compression and optimisation, has signed a strategic memorandum of understanding with Silicon Valley AI infrastructure startup Modular Inc. Modular develops the Mojo programming language and the MAX inference framework and has an enterprise value of $1.6 billion. The companies will work on integrated AI infrastructure from model compression through execution optimisation and expand MAX into multimodal areas such as image and video generation.