Recent challenges in AI infrastructure financing, software frameworks, and hardware innovations reveal a complex interplay shaping the future of AI data centers. As energy costs rise and new technolog...
DeepX has announced the tape-out of its second-generation AI accelerator chip integrating LPDDR5x memory with processing-in-memory technology, promising up to 6.7 times performance gains while operati...
Envision, a Chinese AI infrastructure company, has launched the world’s first 800-volt DC power grid in an AI data center to improve energy efficiency and reduce power losses. This deployment supports...
The recent $87 billion market value loss for Broadcom following Google's partnership with Marvell marks a pivotal shift in AI chip competition. Marvell's move into custom AI inference accelerators, le...
Samsung Electronics announced an expansion of its High Bandwidth Memory base-die production capacity to address supply shortages affecting AI chip manufacturers. The move targets a critical bottleneck...
OpenAI has launched BioBasicsBench, a benchmark with 100 tasks designed to evaluate AI performance in wet-lab scientific tasks, including data interpretation and experimental procedure execution. This...
Dropbox acquired Tigris Data for $350 million to launch the DBX Agent Cloud, a platform tailored for AI developers requiring specialized data infrastructure. This acquisition positions Dropbox to comp...
Poland’s NASK institute has announced plans for the National Data Processing Center, designed to support AI workloads with an initial 15MW power capacity. The center aims to enhance local AI infrastru...
Cerebras Systems’ CS-4 AI compute platform introduces a switchless, wafer-scale architecture that reduces latency and complexity in AI clusters. This design innovation aligns with emerging trends in h...
NVIDIA has launched DSX MaxLPS, a new system designed to enhance throughput and energy efficiency in AI data centers by optimizing power and compute resource management. Early deployments have shown u...


