Friday, April 18, 2025

CISA Last-Minute Extension of Mitre CVE Contract

Nvidia H20 Exports to China Restricted by AI Chip Regulations

Understanding the Investigatory Powers Tribunal

Clear Path Emerges for the Commercial Adoption of Quantum Computing

Hertz Alerts UK Customers to Cleo-Related Data Breach

US Commerce Secretary Anticipates Semiconductor Tariffs Within Two Months

The Significance of Small Language Models in Enterprise AI Applications

Investigator Takes Legal Action Against Police Surveillance and Social Media Monitoring

What’s the Government’s Plan for Digital Verification Services? I’m Curious!

IBM Enhances AI Mainframe Capabilities with Z17

IBM just rolled out its latest mainframe, the Z17, promising to redefine AI at scale. This new machine runs on the IBM Telum II processor and is set up to handle tasks like generative AI and agentic AI.

What makes the Z17 stand out? IBM claims it can process 50% more AI inference operations daily compared to the Z16. With a second-generation AI accelerator baked into the Telum II processor, the Z17 is designed for speed and efficiency. This accelerator boosts frequency and computing power while offering a 40% larger cache. It can handle over 450 billion inference operations in a day, responding in just one millisecond.

Later this year, IBM will introduce the Spyre Accelerator, which will enhance the Z17’s AI capabilities. This combination will enable the system to run multiple AI models, including large language models and AI assistants, all while ensuring security.

Ric Lewis, IBM’s senior vice president for infrastructure, highlighted the importance of solid infrastructure for AI, saying only 1% of enterprise data currently feeds into large language models. With the Z17, businesses can tap into their unused data securely and cost-effectively.

On the AI operations front, IBM has included Watsonx Code Assistant for Z and Watsonx Assistant for Z in the Z17. The latter integrates with Z Operations Unite, offering AI chat-based incident detection and resolution using live data for the first time.

IBM also unveiled IBM Z Operations Unite, which pulls together operational data from various sources in OpenTelemetry format. This tool aims to streamline IT operations using AI, allowing administrators to spot anomalies faster, cutting resolution times from hours to minutes. It correlates events intelligently from multiple sources and collaborates with Watsonx Assistant for Z to offer expert guidance and actionable insights.

In addition, IBM is helping clients transition to quantum-safe encryption. Tina Tarquinio, IBM’s chief product officer, mentioned that while implementing this technology can be challenging, they’re introducing tools to make the process easier for customers.