Skip to Content
Henkel Adhesive Technologies

Henkel Adhesive Technologies

An inside-out view: data centers

The convergence and synergy of performance and sustainability required to power the next generation of AI data centers.
7 min.
Woman in server room data center

The surging demand for AI data centers is transforming the semiconductor industry, redefining the requirements for the chips that power them. According to McKinsey, AI-ready data center capacity is expected to grow by an average of 33% annually from 2023 to 2030. Over the same period, data center power consumption is projected to rise by 160%—a shift largely driven by generative AI. For example, a single ChatGPT query can consume up to ten times more power than a standard Google search.

To support this AI-driven evolution, semiconductor engineers and manufacturers are at the forefront of innovation. They are developing cutting-edge materials, refining fabrication processes, and advancing packaging designs to push the limits of bandwidth, speed, and efficiency—ensuring AI data centers can meet escalating performance demands.

To better understand these shifts, Henkel Adhesive Technologies surveyed semiconductor manufacturers about the evolving landscape of data center chips, the most pressing challenges, and how they are being addressed. Below is a high-level summary of our findings, with a deeper analysis and survey demographics available here.

Pursuing two synergistic priorities: enhanced sustainability and performance

According to survey respondents, the two most pressing and simultaneous semiconductor priorities in the design and manufacture of data center semiconductors are**:

70.80%

Attaining sustainability goals

69.20%

Performance gains in CPUs and GPUs

These longstanding semiconductor priorities are converging, evolving, and shifting in the ongoing pursuit of dual optimization. The good news is that these two trends are synergistic; achieving one helps to bolster the achievement of the other.

Advancing sustainability goals

Sustainability has become a top priority across industries, and semiconductor manufacturing is no exception. However, semiconductor manufacturers have a built-in advantage, thanks to their highly efficient operations, use of renewable energy, and initiatives focused on water conservation and waste reduction. A key focus is maximizing energy efficiency—not just in manufacturing but throughout the entire semiconductor lifecycle and supply chain.

One priority is to reduce material waste in semiconductor manufacturing.

Survey respondent

Smart material selection and strategic supplier partnerships play a crucial role in enhancing chip reliability and longevity. More durable semiconductor packages reduce the need for additional manufacturing, lowering energy consumption and emissions and creating a synergy between performance gains and sustainability. Additionally, reusing and recycling older semiconductor components contribute to the circular economy, further advancing sustainability efforts.

Semiconductors also support energy efficiency in AI data centers. As one survey respondent noted, “Chips are designed to adjust power use based on workloads.” This adaptability helps optimize energy consumption, supporting data center sustainability while maintaining high performance.

Data center in dark with red and blue lighting

Elevating the performance of CPUs/GPUs

AI-driven data centers demand greater bandwidth, speed, and efficiency to handle increasingly intense compute workloads. As a result, achieving significantly higher semiconductor performance is essential to power these advanced infrastructures. The demand for AI-enabled chips is soaring, with market growth nearly doubling between 2023 and 2025, reaching an estimated $92 billion in annual revenue.

Beyond performance, AI’s increasing computational needs introduce additional challenges, including enhanced thermal management and the ability to handle more complex, integrated semiconductor packaging designs. These advancements must be achieved while maintaining strict cost controls and optimizing production efficiency—critical factors in the competitive semiconductor industry.

Survey respondents identified these top data center semiconductor performance requirements**:

Achieve high bandwidth memory 

70.80%

Increase performance for faster network speeds

66.70%

Improve thermal management

64.20%

Manage production cost pressures

62.50%

Handle higher semiconductor integration and sophistication

61.70%

**Percentage of respondent ratings as "very important" or "extremely important"

Adopting material solutions that help

New, AI-enabled semiconductor designs and packaging mandate new material solutions. Highly engineered materials like encapsulation, lid/stiffener attach, and thermal management materials are vital enablers in advanced packaging designs. These materials facilitate the manufacture of complex, delicate structures that can be efficiently processed in high volume at high yield and meet end-use application needs.

Survey respondents identified the following highly engineered material types as essential to the manufacture and end-use performance of semiconductor’s advanced packaging solutions**:

Encapsulation materials

63.30%

Lid stiffener attach materials

62.50%

Underfillers between the die and substrate

60.00%

Wafer-level handling of large, thin dies

59.20%

Thermal management materials

56.70%

**Percentage of respondents rating as “very important” or “extremely important”

Why are these sophisticated materials so essential?

Advanced semiconductor packaging relies on large die configurations that require robust protection, durability, and rigidity to reinforce the package and prevent warpage or cracking—key failure points in advanced devices. Highly engineered materials such as encapsulants, underfillers, and lid/stiffener attach adhesives play a critical role in safeguarding semiconductors during manufacturing and throughout their lifecycle. Beyond protection, these materials are formulated for manufacturing efficiency, featuring application and curing properties tailored for high-speed, high-precision, and high-yield processes, ensuring productivity.

Semiconductor being manufactured by laser

Fueling AI data centers with high-performing, sustainable semiconductors

Driving semiconductor data center performance to new heights while enhancing sustainable practices are key priorities in the next AI-powered era. The convergence of these objectives creates a synergy for semiconductor designers and manufacturers. Tackling these challenges with groundbreaking innovations, advanced engineered materials, and strong industry collaborations will ensure semiconductors continue their legacy of achieving the impossible.

At Henkel, we collaborate with top semiconductor technologists to develop and implement materials that enable sustainable, high-performance advanced packaging designs and meet the growing demands of AI data centers. Contact us to discuss your specific needs.

Related content

Looking for solutions? We can help

Our experts are here to learn more about your needs.

  • A female call-center employee smiling and wearing a headset while working in an office.

    Request a consultation

  • LOCTITE 263 Threadlocker - Application on Bolt

    Request a sample

  • A black female employee scans packages in a warehouse. In the foreground there is the woman with the yellow scanner, in the background scaffolding can be seen.

    Submit an order request

Looking for more support options?

Our support center and experts are ready to help you find solutions for your business needs.