GPUs Topic - AI Updates

Investment capital flows through AI laboratories and data centers toward high-performance processors in a circular financial system

Nvidia has committed nearly $50 billion to AI laboratories that buy its chips while helping financing partners target more than $500 billion for infrastructure.

Millions of next-generation AI processors flow into a large cloud data center and connected robotics operation

View Post

Chips & Semiconductors

Amazon Adds Two Million Nvidia GPUs to Its AI Infrastructure Plans

AWS will deploy two million more Nvidia GPUs in 2027 and 2028 while extending the partnership into networking, models, robotics and edge computing.

An engineer examining a GPU server while streamlined light patterns suggest faster AI inference

View Post

Compute & Infrastructure

Kog Wants Software—not New Chips—to Unlock Faster AI Inference

French startup Kog claims deep GPU optimisation can accelerate single-request AI inference, but its most dramatic benchmark still relies on a purpose-built small model.

Technology leaders mapping a constrained supply chain from memory chips and accelerators to powered data centre racks

View Post

Chips & Semiconductors

AI Chip Shortages Turn CTO Roadmaps Into Supply-Chain Plans

GPU, memory, packaging, power, and geopolitics stretched AI deployments in 2025, forcing technology leaders to design for scarcity rather than vendor promises.

Distributed computing providers contributing spare GPU capacity to a transparent shared marketplace

View Post

Compute & Infrastructure

Five Compute Marketplaces Turn Idle GPUs Into AI Capacity

Argentum, Aethir, Bittensor, Akash, and Flux use different decentralised mechanisms to match AI demand with underused computing resources.

A secure enterprise data core connected to a vast GPU cloud computing environment

View Post

Compute & Infrastructure

Oracle and NVIDIA Bring Enterprise AI Closer to the Data

Oracle and NVIDIA are joining immense GPU clusters with in-database agents, accelerated data tools, and simpler deployment to make enterprise AI practical at scale.

Large dual-die computing package linked to liquid-cooled accelerator modules

View Post

Chips & Semiconductors

Nvidia Blackwell Targets Trillion-Parameter Real-Time AI

Blackwell combines a 208-billion-transistor GPU, faster interconnects and a new Transformer Engine to reduce the cost and energy of very large AI workloads.