Search AI Products and News

French AI startup Mistral AI announced on Wednesday its full entry into the AI infrastructure sector, positioning itself as a strong response to American cloud giants from Europe. At the same time, the company also launched a new inference model that can rival the most advanced systems of OpenAI. This Paris-based company released Mistral Compute, an integrated AI infrastructure platform developed in collaboration with NVIDIA (Nvidia), designed to provide European enterprises and governments with alternative solutions to摆脱对

At the recent VivaTech科技Summit held in Paris, NVIDIA CEO Jensen Huang stated that European artificial intelligence (AI) computing power is expected to increase tenfold within the next two years. This prediction has drawn significant attention from the industry, especially given the increasing worldwide focus on AI technology. Huang pointed out that Europe is actively building more than 20 "AI superfactories", which will provide strong infrastructure support for AI development. He emphasized that countries in Europe have already realized

Mark Stevens (NVIDIA Corp.) sold over one million shares of the company's stock this week, with the total transaction value approaching $150 million. This move comes after NVIDIA's stock experienced some volatility and has seen a recent recovery in its price. According to documents released by the U.S. Securities and Exchange Commission (SEC) on Wednesday, Stevens' stock sales were conducted in two separate transactions on Monday and Tuesday. As NVIDIA's performance in the market gradually improves...

NVIDIA officially released Llama Nemotron Nano VL on June 3, 2025. This compact visual-language model (VLM) is specially optimized for document intelligence processing. The model topped the OCRBench v2 benchmark test, showcasing its outstanding capabilities in handling complex documents, charts, and video frames. With efficient inference performance and flexible deployment options, Llama Nemotron Nano VL provides enterprises with solutions from the cloud.

Recently, tech giant NVIDIA, in collaboration with the Massachusetts Institute of Technology (MIT) and the University of Hong Kong, released a new framework called Fast-dLLM. This innovative framework aims to significantly increase the inference speed of diffusion models (Diffusion-based LLMs), up to 27.6 times, providing stronger technical support for AI applications. The challenges and opportunities of diffusion models make them strong competitors to traditional autoregressive models.

According to the Financial Times, Oracle will spend approximately $40 billion to purchase Nvidia's latest superchips, with plans to provide computing power for OpenAI. These superchips will be deployed in the first US "StarGate" data center located in Abilene, Texas. However, there are still questions about whether this data center can provide sufficient power to meet such massive computing demands. Image source note: The image is generated by AI and provided by MidJourney. This large investment will be used to purchase

The global leader in open-source solutions, Red Hat, recently announced the launch of the revolutionary open-source project LLM-D, specifically addressing the urgent needs for large-scale inference in generative AI. The project, founded by industry giants such as CoreWeave, Google Cloud, IBM Research, and NVIDIA, aims to meet the most stringent production service-level goals through breakthrough technologies that enable cloud-based large language model inferencing.

According to Reuters, citing three sources familiar with the matter, Nvidia will launch a new AI chip designed specifically for the Chinese market in June. The pricing is expected to be between $6,500 and $8,000, which is significantly lower than the current H20 series priced around $10,000 to $12,000. The new chip is based on the latest Blackwell architecture and adopts the RTX Pro6000D design framework. To bypass US export restrictions on advanced chips to China, this GPU deliberately avoids TSMC's advanced packaging processes and instead uses standard GDDR7.