Section
Infrastructure – Page 14
The latest Infrastructure news, analysed daily by the i10X newsroom.
FilterCompanies & models
More stories
-
PrfaaS: Global KV Cache for Efficient LLM Serving
Discover PrfaaS, a groundbreaking architecture from Moonshot AI and Tsinghua University that distributes the KV Cache across datacenters worldwide, reducing costs and latency for global AI…
-
AI Agent Long-Term Memory: Trends and Challenges
Discover the shift to universal long-term memory for AI agents using Mem0 and ChromaDB. Learn how this architecture enables persistent context, boosts agent reliability, and addresses…
-
Anthropic Partners with CoreWeave: AI Infrastructure Shift
Explore Anthropic’s strategic partnership with CoreWeave, adopting a multi-cloud approach to secure GPU resources and optimize AI model performance. Learn why this matters for AI labs…
-
LFM-2.5-VL-450M: Compact On-Device VLM for Edge AI
Discover LFM-2.5-VL-450M, Liquid AI’s 450M-parameter vision-language model for on-device use. It fuses multilingual understanding with object localization for sub-250ms inference, enabling privacy-focused, real-time AI apps. Explore…
-
Broadcom-Google-Anthropic AI Deal: 3.5 GW TPU Compute
Explore the strategic Broadcom, Google, and Anthropic collaboration providing 3.5 GW of TPU capacity for AI training. This vertically integrated stack challenges NVIDIA dominance and secures…
-
Data Center Efficiency: Beyond PUE to Software Optimization
Explore the shift in data center efficiency from traditional PUE metrics to holistic software-defined approaches. Optimize IT workloads and flash storage to cut energy waste and…
-
Rising AI Compute Costs: Shift to Efficiency
As AI labs face massive compute bills, the industry pivots from unchecked growth to economic efficiency. Discover how optimizing TCO for training and inference is the…
-
Google’s Gemma 4: Open AI Models for On-Device Deployment
Explore Google’s Gemma 4 family of open AI models designed for seamless deployment from mobile devices to cloud infrastructure. Optimized for efficiency, privacy, and low-latency applications.…
-
AWS Agentic AI Architecture: Reliable Production Build
Explore AWS’s new reference architecture for building production-grade agentic AI using Bedrock AgentCore and Nova Sonic 2.0. Overcome open-source challenges with reliable orchestration and low-latency responses.…
-
NVIDIA PivotRL: 4x Efficient Training for Agentic AI
Discover NVIDIA’s PivotRL framework, a reinforcement learning breakthrough that cuts AI agent training costs by 4x through fewer simulations. Ideal for complex tasks like software engineering.…
-
OpenAI Pivots to Hyperscalers for AI Compute Power
OpenAI is shifting from building its own data centers to relying on Microsoft Azure and Oracle Cloud for massive GPU capacity. Explore the implications for AI…
-
Rebuilding Deep Q-Learning on JAX: RL Tooling Shift
Explore the shift from PyTorch and TensorFlow to JAX for implementing Deep Q-Learning. Discover how this functional paradigm boosts performance, scalability, and efficiency in reinforcement learning.…