AI on IBM z17, Meta’s Llama 4 and Google Cloud Next 2025

Summary: The video discusses the insights into unstructured data in enterprises, featuring a conversation among industry experts about the percentage of this data, the launch of IBM Z, and the latest developments in artificial intelligence, particularly in relation to data security and efficiency. It also touches on the mix of structured and unstructured data, the growing role of AI in business, and the significance of recent AI model releases.

Keypoints:

• Estimates suggest that approximately 90% of enterprise data is unstructured, highlighting the challenge of data management.
• Guests include Kate Soul, Show Varsn, and Hillary Hunter, discussing the importance of mainframes in financial transactions.
• IBM’s Z system boasts eight nines of reliability and integrates AI to enhance transaction processing and fraud detection.
• Multi-model AI on IBM’s mainframe can handle billions of transactions daily, scoring transactions in real-time.
• Significant improvements in the efficiency of AI processing on mainframes, with a 17% reduction in power consumption noted.
• The launch of Meta’s Llama 4, featuring large language models, introduces more powerful open-source architectures to the community.
• Industry leaders discuss evolving strategies in AI deployment while considering data privacy and sovereignty concerns.
• The importance of adapting AI models to enterprise-specific needs and the potential for innovation in the space.
• Google Cloud’s announcements regarding AI on-premises solutions affirm the necessity for AI deployment in secure environments.
• The Pew Research report emphasizes a disconnect between expert optimism about AI and public concerns regarding job impacts, highlighting a need for diverse perspectives in AI research.