infrastructureScore 55/100Watch

Windows GPU integration and confidential computing on Vera Rubin requires encrypted workings memory and encrypted data in transit/use, driving security infrastructure demand

Jensen Huang· NVIDIA· AI· 2024-05-21· about Microsoft, NVIDIA (MSFT, NVDA)
The entire coding path from storage, which is the long-term memory, the working memory is encrypted, the data is encrypted in transit, the data is also encrypted in use. So we're going to really innovate in the area of confidential computing.

Why it matters

Encrypted execution becomes mandatory for agentic systems handling sensitive data, creating new infrastructure requirements and certification standards.

Investment implication

Security hardware vendors, cryptographic accelerator makers, and secure enclave software providers gain critical mass. Data center operators must upgrade security stack.

Source

Jensen Huang and Satya Nadella's Conversation at Microsoft Build (YouTube)
← All signals