$NVDA (-1.79%)
$IOS (+1.59%)
$SAP (-2.67%)
$SIE (-1.64%)
$DTE (-1.83%)
My dears,
while one CEO sits in the sandbox in a huff. The brilliant Jensen Huang goes on a trip to Europe.
Jensen Huang is coming to Berlin next week to meet German Chancellor Friedrich Merz. The CDU/CSU and SPD want to promote an AI gigafactory in Germany with 100,000 GPUs.
At the end of next week, Federal Chancellor Friedrich Merz (CDU) wants to meet Jensen Huang, the head of the US GPU manufacturer Nvidia, in Berlin. Handelsblatt has learned this from government and industry circles. A meeting with Vice-Chancellor Lars Klingbeil (SPD) is also reportedly planned.
The CDU/CSU and SPD have agreed to build an AI gigafactory in Germany, which is to be powered by 100,000 GPUs.
Nvidia boss Huang is traveling through Europe for most of the week. On Monday, he will first appear at Tech Week in London with UK Minister for Investment Poppy Gustafsson. On Tuesday, he is scheduled to give a keynote speech at the Vivatech technology conference in Paris. According to estimates in industry circles, Huang is likely to announce various initiatives relating to the development of computing infrastructure in Europe during his appearances.
The EU has announced a funding program for the construction of AI Gigafactories, which are suitable for training large language models. In each of the four to five planned facilities, 100,000 or more special data center GPUs are to be used. Nvidia is the market leader here.
SAP, Deutsche Telekom, Ionos and Siemens are planning a joint application to the European Union (EU) for the AI data centers.
Why GPUs are so important for AI and artificial neural networks
GPUs are versatile and ideally suited for parallel computing, as is often the case in AI - but they are also energy-intensive and expensive, reported our author Klaus Manhart. They are most frequently used for training AI models. As this is very computationally intensive, several GPUs are connected together so that they can all train an AI system synchronously.
AI-optimized processors are available from almost all major processor providers. Nvidia - known for its GPUs - has entered the high-end AI market with its Tensor Core GPUs. The core GPUs are an integral part of their AI-optimized GPU architecture and are designed to perform and aggregate matrix multiplications and complex tensor operations as quickly as possible. The latest models accelerate training and inference of large neural networks in data centers.
https://www.golem.de/news/ai-gigafactory-kanzler-merz-trifft-nvidia-chef-in-berlin-2506-196943.html
