The centralized mega-cluster model for AI infrastructure is reaching its limits due to physical, social, and environmental constraints. Distributed AI inference architectures, composed of specialized ...
d-Matrix and GigaIO announced a partnership on April 3, 2026, to develop scalable AI inference infrastructure combining d-Matrix’s AI software with GigaIO’s composable fabric technology. The collabora...
A startup founded by a former Goldman Sachs banker has secured multi-million dollar funding led by SoftBank to accelerate the development and production of AI chips optimized for performance and effic...
D-Matrix has acquired GigaIO’s data center business to integrate advanced interconnect technology into its rack-scale AI infrastructure. This acquisition aims to improve AI inference performance and s...
TeraWulf has sold its stake in a nuclear-powered facility to pivot from Bitcoin mining to developing AI data centers powered by sustainable nuclear energy. This strategic move addresses the rising ene...
Mistral, a French AI startup, has raised $830 million to build a data center near Paris equipped with thousands of Nvidia GPUs. This facility aims to enhance Europe's AI cloud infrastructure, ensuring...
CoreWeave announced a strategic pivot on April 2, 2026, to focus solely on AI inference workloads, partnering with AMD to optimize its GPU cloud services for power-efficient, low-latency AI model depl...
CoreWeave announced a major investment in March 2026 to expand its AI inference infrastructure, focusing on specialized GPUs and hardware optimized for real-time AI applications. The company aims to c...
Rambus announced its HBM4E memory controller in March 2026, featuring a 16 GT/s speed and a 2,048-bit interface that supports up to 4 TB/s bandwidth. Targeting AI and HPC workloads, this controller ai...
Marvell announced an expansion of its networking solutions to enhance support for Nvidia's AI ecosystem, focusing on higher bandwidth and lower latency critical for AI workloads. The upgrades target h...