TL;DR:
- Fastweb acquires thirty-one NVIDIA DGX H100 systems for Lombardy data centers.
- Utilizing NVIDIA DGX SuperPOD technology with 248 NVIDIA H100 Tensor Core GPUs.
- Fastweb plans to offer AI infrastructure in the cloud for AI and generative AI applications.
- Development of the first Italian-language large language model (LLM) using NVIDIA AI Enterprise software.
- LLM is tailored to Italian language nuances, grammar, context, and culture.
- Integration of NVIDIA DGX H100, LLM, and Fastweb’s cloud infrastructure for generative AI application development.
Main AI News:
Fastweb, a prominent player in the telecommunications arena, is set to usher in a new era of technological innovation in Italy. In a strategic move, the company has procured thirty-one cutting-edge NVIDIA DGX H100 systems, earmarked for installation in state-of-the-art data centers located in Lombardy, Italy. This sophisticated system, boasting the latest NVIDIA DGX SuperPOD reference architecture and equipped with a formidable arsenal of 248 NVIDIA H100 Tensor Core GPUs, is slated to become operational by the first half of 2024.
The significance of this substantial investment lies in its potential to empower Fastweb to leverage the unparalleled capabilities of the NVIDIA DGX H100 cluster for artificial intelligence (AI) endeavors. Fastweb aims to offer this formidable AI infrastructure to its customers via the cloud, as an Infrastructure-as-a-Service (IaaS) solution, thereby facilitating the development of AI and generative AI applications.
Fastweb’s ambitious vision extends to harnessing the prowess of NVIDIA AI Enterprise software to forge a pioneering achievement – the creation and deployment of the first large language model (LLM) natively trained in the Italian language. An assembly of proficient generative AI experts will delve deep into the realm of deep learning systems, meticulously crafting this LLM using vast repositories of Italian language data. This groundbreaking LLM will serve as a platform uniquely tailored to Italy, serving both Fastweb and third-party developers in their quest to construct a diverse array of generative AI applications. It will possess the capability to capture the nuances of the Italian language, including its intricate grammar, context, and cultural identity.
The convergence of the potent NVIDIA DGX H100 systems, the indigenous LLM, and Fastweb’s robust cloud and cybersecurity infrastructure promises to provide a comprehensive ecosystem. Within this ecosystem, enterprises, public administrations, startups, and various stakeholders can engage in the development of generative AI applications across diverse industry verticals.
Ronnie Vasishta, Senior Vice President of Telecom at NVIDIA, emphasizes the transformative nature of this endeavor, stating, “As the industry pivots towards cloud computing, the telcos of the future will not merely serve as connectivity providers. Fastweb’s commitment to Italy’s digital transformation marks a significant stride in delivering robust, efficient, and locally attuned computing resources to AI pioneers in the nation.“
Walter Renna, CEO of Fastweb, echoes this sentiment, affirming that through this strategic investment, Fastweb reaffirms its role as a catalyst in Italy’s digital evolution. This initiative lays the foundation for developing AI-based services with a national and sustainable approach, ushering in an era where Fastweb will provide high-performance, transparent, and explainable systems to enterprise clients, public entities, and the broader Italian ecosystem for the advancement of generative AI applications.
Conclusion:
Fastweb’s strategic investment in NVIDIA DGX technology and the development of an Italian-language large language model positions them as a major player in Italy’s digital transformation. This move signifies a significant shift in the telecommunications market, where telcos are evolving into providers of advanced AI infrastructure and services, paving the way for innovation across various industries in Italy.
