AI Next Generation Data Center
"Smaller but very Powerful!"
We are an expert team designing the next-generation AI data centers in Europe (Spain, Portugal, France and Switzerland).
Our parent companies (Penguin Solutions & SK Hynix) are world top players in AI related industries:
www.penguinsolutions.com
www.skhynix.com
The data centers we are building are specially adapted for AI/machine learning, blockchain, and other high-power computing workloads to run on the massive array of GPUs integraged with a pool of high bandwidth memory modules, high-density storage arrays, and an ultra-low latency fiber-optic network fabric.
One of the key features of the data centers we design is the full adoption of liquid immersion cooling systems which allow us to achive the highest heat dissipation capacity and the lowest PUE, without getting bound by the inefficiencies of legacy data center infrastructure.
Our approach is to integrate energy balanced and storage solutions and to select specific optimized computing AI systems in totally immersive cooling modular pods that can dramatically preserve the capacity of superchips multiprocessing tasks while keeping low operation costs and construction footprint. The environmental impact compared with conventional data centers is significantly improved.
concept design
Key Design Features
Powered Shell & Colocation
Our generation AI Data Centers will provide powered shell as well as wholesale/retail colocation services. Powered shell services gives a tenant total control over the configuration of its dedicated space under the constraint of power budget. Whoesale/retrail colocation service allows a tenant to use one of more immersion cooling pods (or a fraction of an immersion cooling pod) in which its own server platforms, storage devices, and network switches can be installed and operated.
Sovereign Cloud Platform
Concerning our Data Centers located in Switzerland, we will also offer Swiss-sovereign cloud services which will give our clients the assurance of tight protection of the data privacy under the strong Swiss data privacy law and the composability of infrastructure and platform resources (especially GPUs) on demand.
Why Building AI-specific data centers, as opposed to using standard data centers, is essential ?
1) Massive Computing Power: AI applications, especially those involving machine learning and deep learning, require significant computational resources. Data centers provide the necessary infrastructure to handle these intensive tasks by hosting specialized hardware like GPUs and TPUs.
2) Data Storage and Processing: AI systems need to process and analyze vast amounts of data. Data centers offer the storage capacity and processing power to manage these large datasets efficiently.
3) Scalability: As AI technologies evolve, the demand for computational resources grows. Data centers can scale up to meet these increasing demands, ensuring that AI applications can continue to develop and improve.
4) Efficiency and Optimization: AI can help optimize data center operations by predicting and allocating workloads more efficiently. This reduces energy consumption and improves overall performance.
5) Support for Innovation: With AI driving advancements in various fields, from healthcare, education, finance, robotics to smart cities, having dedicated data centers ensures that these innovations can be supported and scaled effectively.
In essence, AI data centers are the backbone of the AI ecosystem, providing the necessary infrastructure to support the development and deployment of AI technologies.