NVIDIA Introduces NVFP4 for Efficient 4-Bit AI Model Pretraining
NVIDIA has unveiled NVFP4, a 4-bit format designed to enhance AI model pretraining efficiency, offering the precision of 16-bit with the speed of 4-bit.
Advancements in semiconductor technology, GPU development, and data center infrastructure that power AI computing capabilities.
NVIDIA has unveiled NVFP4, a 4-bit format designed to enhance AI model pretraining efficiency, offering the precision of 16-bit with the speed of 4-bit.
China is integrating AI into its economy with a national computing network, aiming to add trillions to its GDP by 2035.
Lambda has deployed Supermicro GPU-optimized servers, including NVIDIA Blackwell systems, to enhance its AI infrastructure, announced in a press release. This collaboration aims to deliver high-performance AI solutions at scale.
Axelera AI, backed by Samsung and the European Union, is in talks to raise over €150 million to enhance its edge AI capabilities, according to Bloomberg.
Fermi America has announced a partnership with Parkhill and Lee Lewis Construction to build an 11-gigawatt private AI grid in Amarillo, Texas, according to a press release.
Meta Platforms has signed a deal with Silicon Ranch to develop a 100MW solar farm in South Carolina, powering its upcoming AI data center entirely with renewable energy.
Meta Platforms has entered a six-year cloud computing agreement with Google Cloud, valued at over $10 billion, according to Reuters. This deal will support Meta's AI infrastructure needs.
Crusoe, the developer of OpenAI's first Stargate data center, is in talks to raise $1 billion at a $10 billion valuation, according to The Information. The company plans to expand its cloud services by acquiring Atero, a Tel Aviv-based startup.
BAE Systems is accelerating chip production for defense programs by modernizing its Nashua facility with CHIPS Act funding.
Blackstone Credit & Insurance has increased its debt facility for Aligned Data Centers to over $1 billion, supporting the company's expansion plans across the Americas.
Chinese AI startup DeepSeek has released an upgraded version of its V3 model, optimized for Chinese-made chips and featuring enhanced processing speeds.
OpenAI is exploring the possibility of selling AI infrastructure services to other companies, as stated by CFO Sarah Friar. This move could create a new revenue stream to offset the company's high operational costs.
GI Partners has announced a $1 billion investment in Flexential to accelerate the development of next-generation data centers across the U.S., according to a press release.
Lightning AI has introduced a Multi-Cloud GPU Marketplace, offering AI teams access to GPUs from leading cloud providers, potentially reducing AI costs by 70%.
Yondr Group has announced the acquisition of a 163-acre site in Lancaster, Texas, to develop a 550MW data center campus, marking a significant step in its US expansion.
AVAIO Digital has announced plans to construct a $6 billion AI-ready data center hub in Brandon, Mississippi, set to be completed by 2027.
The Generative AI Data Center Market is projected to grow from $6.584 billion in 2024 to $19.9 billion by 2031, driven by global AI adoption and digital transformation, according to Valuates Reports.
Elea Data Centers has signed a Memorandum of Understanding with Oracle and the City of Rio de Janeiro to advance the Rio AI City project, a major data center hub in the Olympic Park.
Oracle has integrated OpenAI's GPT-5 into its database and cloud applications, enhancing AI capabilities for enterprise customers, announced in a press release.
Applied Digital is set to build a 280MW AI data center campus near Fargo, North Dakota, with construction starting in September 2025, according to Data Center Dynamics.