AI Factories: What do you need to know?

Artificial Intelligence (AI) is not just a technological trend—it’s the backbone of modern innovation. From autonomous driving to generative AI, the processing power and infrastructure needed to train, operate, and deploy AI models at scale are growing at an unprecedented pace. At the forefront of this transformation are AI factories—large-scale, high-performance data centers optimized for AI workloads.

At the NVIDIA GTC 2025 conference, CEO Jensen Huang declared that "every company will eventually have two factories: one for their products and one for AI." This statement underscores the critical role AI factories will play in the future of business and technological innovation. In Israel, MedOne, the country’s leading data center provider, is positioning itself as the hub for AI infrastructure, aligning closely with NVIDIA's AI factory vision.

What Are AI Factories?

An AI factory is essentially a specialized data center designed to handle the computational and data processing needs of AI models at scale. Unlike traditional data centers, AI factories are built with high-performance GPUs, optimized networking, and massive data storage to enable training and inference for complex AI models.

AI factories are capable of supporting a range of AI functions, including:

• Model Training – Processing large datasets to train deep learning models.
• Inference – Running trained models in real-time to deliver insights and predictions.
• Data Processing – Managing large-scale structured and unstructured data to enable AI development.
• Multi-Model Operations – Supporting diverse AI models simultaneously.

NVIDIA's AI factory vision extends beyond just building the infrastructure—it includes software and systems that enhance the entire AI pipeline. This vision is driving the next generation of AI capabilities across industries, from healthcare and finance to autonomous vehicles and manufacturing.

NVIDIA’s AI Factory Platform: Powering the Future of AI

At GTC 2025, NVIDIA unveiled a suite of products and platforms designed to accelerate the development of AI factories:

1. Blackwell Ultra GPU Platform

The Blackwell Ultra GPU, set to launch in late 2025, is designed to deliver up to 50% higher performance than previous generations. This enhanced computational power is critical for AI model training, particularly for large language models (LLMs) and generative AI applications.

2. NVIDIA Rubin AI Superchip

The Rubin AI Superchip platform is designed to integrate AI training and inference within a single system. It offers increased memory bandwidth and enhanced power efficiency, making it ideal for real-time AI processing.

3. Omniverse for AI Factories

NVIDIA's Omniverse platform allows businesses to simulate and design AI factories before they are built. This "digital twin" approach enables companies to optimize the design, configuration, and operation of AI infrastructure, reducing costs and increasing efficiency.

4. NVIDIA Dynamo

Dynamo is an open-source inference software platform designed to scale AI models efficiently. It allows AI factories to deploy and run complex models with reduced latency and increased resource utilization.

MedOne: Israel's AI Factory Pioneer

In Israel, MedOne has established itself as the backbone of the country’s digital infrastructure. Founded in 1997, MedOne operates over 25,000 square meters of underground data center capacity across four highly secure facilities. With seven more facilities under development, MedOne is actively expanding to meet the growing demand for AI-driven data services.

Strategic Expansion for AI Workloads

MedOne has announced an investment of approximately $270 million to build two new underground data centers near Tel Aviv, adding an additional 30,000 square meters of capacity. These facilities will be designed specifically to accommodate AI factory infrastructure, including:

• High-performance GPUs for AI training and inference.
• Direct connectivity to global cloud providers and networks.
• Advanced cooling systems to handle the heat generated by AI processing.

These new facilities will position MedOne as the central hub for AI development and deployment in Israel, enabling companies to leverage AI for competitive advantage.

Strategic Partnerships and Ownership

MedOne’s success in AI infrastructure is strengthened by its robust ownership structure and strategic partnerships:

• Avraham Livnat Ltd – A major industrial and infrastructure player in Israel, providing financial backing and strategic insight.
• Taavura Holdings Ltd – Israel's largest logistics company, offering operational expertise and supply chain support.
• Berkshire Partners – A leading U.S.-based private equity firm specializing in infrastructure investments. Berkshire acquired a 49% stake in MedOne in 2022, valuing the company at over $430 million.
• van Rooyen Group – A global data center investment firm with expertise in high-performance infrastructure projects.

This diversified ownership base ensures that MedOne has the capital and strategic guidance necessary to scale its AI factory operations rapidly.

Security and Compliance

AI factories demand not only computational performance but also a high level of security and regulatory compliance. MedOne’s data centers are among the most secure in the world, with facilities located underground to protect against natural disasters and physical threats.

Key Security Measures:

• 72-hour autonomous operation – MedOne’s facilities are designed to function independently for up to three days without external power or support.
• ISO 27001 certification – MedOne’s data centers meet the highest international standards for information security management.
• Advanced Access Controls – Biometric and multi-factor authentication systems ensure that only authorized personnel can access critical infrastructure.
• Cybersecurity Monitoring – Real-time threat detection and response systems protect client data from cyberattacks.

These measures make MedOne a trusted partner for financial institutions, government agencies, and multinational corporations that require top-tier data security.

Client Base and Industry Focus

MedOne serves a diverse client base across key industries:

• 1. Financial Institutions
• Banks and financial services rely on MedOne's infrastructure to ensure secure transactions, data integrity, and uninterrupted service.
• 2. Government Ministries
• Israel’s government agencies depend on MedOne for secure, high-availability services to maintain essential public functions.

• 3. Technology Companies
• MedOne provides high-performance computing environments for AI startups and multinational tech giants, enabling them to train and deploy AI models faster.
• 4. Healthcare and Life Sciences

AI is playing an increasingly critical role in drug discovery, medical imaging, and diagnostics. MedOne’s AI factory infrastructure enables healthcare organizations to process large datasets and develop AI-driven solutions.

The Future of AI Factories

AI factories are set to become the new standard for data processing and model development. As AI models become larger and more complex, the need for high-performance, scalable infrastructure will only increase.

NVIDIA’s AI factory ecosystem provides the technological foundation for this shift, while MedOne’s strategic investments and infrastructure position Israel as a leader in AI development.

The next decade will likely see a rapid expansion of AI factories worldwide, with MedOne and NVIDIA at the forefront of this transformation. Companies that invest in AI infrastructure today will be well-positioned to lead the next wave of technological innovation.

Conclusion

AI factories are not just a technological advancement—they are the future of business. NVIDIA’s pioneering hardware and software platforms, combined with MedOne’s secure and scalable infrastructure, create the ideal environment for AI innovation. As AI continues to evolve, MedOne’s strategic positioning in Israel’s tech ecosystem ensures that it will remain a key player in the global AI factory revolution.