Keeping pace through closer customer collaboration

Infrastructure providers for AI data centers must improve energy efficiency and optimize thermal management in increasingly dense systems while adapting to quickly evolving needs. Close dialogue with component suppliers helps align these needs with technological development. Aline Vérité, Global Sales Development for AI Data Centers at TDK, explains how customer roadmaps inspire her work and how TDK’s technologies can contribute.

What makes working with AI infrastructure customers different from more established markets?

Aline Vérité: Speed is key. Development cycles are short, and customer requirements evolve quickly. It is crucial that I understand the challenges customers face today and anticipate future changes to their platforms.

Therefore, I continuously engage in dialogue with them. As customers ramp up their computing capacity, they must also consider power demands, cooling needs, reliability, and available space. These requirements are deeply intertwined, so it is essential for me to understand the bigger picture to select the right component technologies.

How do you identify which customer needs influence product development?

It all starts with listening to the customers and discussing their technical challenges and future roadmaps. These conversations help me determine which of our existing technologies can contribute and where new ones may be needed.

I also monitor market developments and technology trends. Combining these insights helps me guide product development and keep TDK’s priorities aligned with changing requirements. The important point for me is maintaining that exchange as customer plans evolve. Just one conversation cannot capture a market that is moving this quickly.

What does your role mean for customers whose challenges span several technology areas?

I connect emerging customer needs with TDK’s technologies. AI data centers comprise power conversion, passive components, thermal management, and sensing technologies. While customers may know the company for one part of the broad portfolio — particularly capacitors — their requirements extend across several areas.

TDK supports the hardware infrastructure along the entire power chain, from the electricity grid to the processor chip. Thanks to the breadth of our technologies and materials expertise, we can contribute in many ways. We create value for customers by relating these capabilities to their system requirements, whether the issue is power delivery, limited board space, or monitoring operating conditions.
 

The infrastructure of an AI data center consists of various subsystems. TDK contributes to the subsystems highlighted in blue.

Could you provide some examples of how this broader perspective helps to address customer challenges?

One example is power architecture. As customers consider higher-voltage DC distribution, the requirements for capacitors, magnetic components, and protection devices may change. Today, our ultra-compact aluminum electrolytic capacitors are used in server power supply units. However, in other architectures, their role and location may change, including their use in capacitor bank units. High-voltage contactors can provide galvanic isolation when open, depending on the system’s electrical and protection requirements.

At the board level, the challenge is delivering power in very limited spaces. TDK’s µPOL point-of-load converters offer a compact design, high integration, and efficient power conversion. These modules help designers reduce the board space needed for power delivery and simplify power design.

Another practical example is cooling-system monitoring. AI data centers need to operate 24/7. Temperature, pressure, and coolant flow measurements help operators understand how a liquid cooling system is performing. Any changes in these readings may indicate issues that require further examination. When used in monitoring and maintenance strategies, this information can reduce the risk of unplanned downtime.

While each example begins with a specific customer need, they all contribute to the same objective: operating increasingly powerful AI infrastructure efficiently and reliably.

TDK supplies key components for the infrastructure of AI data centers.

How can you support customers whose architecture choices are still evolving?

My goal is to support various architectural choices as requirements evolve. The need to efficiently convert power and monitor critical operating conditions remains fundamental, even when the location or implementation of these functions change.

I discuss these changes with customers and follow emerging trends, such as solid-state transformers and highly integrated power conversion. Component requirements depend on the particular architecture, so it is important for me to understand where a technology fits and what constraints it must meet.

This is why customer roadmaps are so important to me. They help me link technology development to the platforms that customers are planning and determine where TDK’s portfolio needs to evolve.

What customer needs will shape your priorities in the coming years?

Efficient power delivery will remain indispensable. As system density increases, there will be greater demand for compact components and cooling-system monitoring. As infrastructure becomes more complex, customers will require technologies that help them manage these demands while ensuring reliable operation.

My priority is strengthening customer engagement and keeping our development focus closely aligned with their needs. I must understand the practical constraints customers face, bring relevant TDK technologies into the discussion, and maintain that dialogue as they develop their next-generation platforms.