- Nvidia and Meta will jointly optimize workflows for Muse models and agents
- Goal: better training and inference performance, lower total cost of ownership
- Companies cite frontier-level benchmark performance and surging downloads since launch
- Meta to expand Vera Rubin platform and Vera CPU deployment
- Also Spectrum-X Ethernet networking and BlueField-4
- Aimed at next-generation AI agents, maximum performance per watt
- Exploring Nvidia OpenShell reference architecture for safer agent deployment
- Exploring Nvidia Sentry to monitor agent behaviour
- Builds on the February multi-year infrastructure partnership
- Nvidia to invest $3 billion in Meta's equity offering
- Meta raised its offering to $97.5 billion, its largest fundraising
- Nvidia, AMD and Broadcom showed up to $5 billion combined interest in Class A shares
- Analysts flagged questions over circular capital flows in AI
Useful?