The convergence and synergy of performance and sustainability required to power the next generation of AI data centers
The surging demand for AI data centers is transforming the semiconductor industry, redefining chip design and manufacturing priorities. According to McKinsey, AI-ready data center capacity is expected to grow at a 33% annual rate from 2023 to 2030, while data center power consumption could rise by 160%—largely fueled by generative AI workloads. A single ChatGPT query, for example, can consume up to ten times more energy than a standard Google search.
To meet these demands, semiconductor engineers are pushing innovation forward—developing cutting-edge materials, refining fabrication processes, and advancing packaging designs that enhance bandwidth, speed, and energy efficiency.
To better understand these industry shifts, Henkel Adhesive Technologies surveyed semiconductor manufacturers about their biggest challenges and evolving strategies. Below is a summary of key findings.
Pursuing Two Synergistic Priorities: Enhanced Sustainability and Performance
When asked about their top priorities for next-generation data center chips, respondents emphasized two key areas:
- Attaining sustainability goals (70.8%)
- Performance gains in CPUs and GPUs (69.2%)
Notably, these priorities are increasingly seen as mutually reinforcing. Gains in efficiency often simultaneously advance performance objectives, creating a powerful synergy between sustainability and innovation.
“To support sustainable data centers, our focus is power-efficient semiconductor designs.” – Survey respondent
“Low latency memory interfaces are a priority for handling expanding workloads.” – Survey respondent
Advancing Sustainability Goals
Sustainability is now central to semiconductor manufacturing strategies. Companies are reducing environmental impacts by improving energy efficiency, using more renewable energy, conserving water, and minimizing waste throughout the product lifecycle.
“One priority is to reduce material waste in semiconductor manufacturing.” – Survey respondent
Smart material selection and more robust chip packaging are important drivers. Durable semiconductor packages extend product lifetimes, lower replacement needs, and reduce the overall carbon footprint. Companies are also increasingly participating in circular economy efforts by recycling and reusing semiconductor components where possible.
Moreover, semiconductors are being designed to adjust power consumption based on real-time workloads, optimizing energy use in AI-driven data centers without compromising performance.
“Chips are designed to adjust power use based on workloads.” – Survey respondent
Elevating the performance of CPUs/GPUs
As AI workloads become increasingly complex, demand for high-performance semiconductors has never been greater. The market for AI-enabled chips is forecasted to nearly double between 2023 and 2025, reaching $92 billion in annual revenue.
Survey respondents identified the following performance imperatives:
- Achieve high bandwidth memory (70.8%)
- Increase performance for faster network speeds (66.7%)
- Improve thermal management (64.2%)
- Manage production cost pressures (62.5%)
- Handle higher semiconductor integration and sophistication (61.7%)
Advanced packaging approaches like 2.5D/3D integration and heterogeneous architectures are becoming mainstream solutions. However, these complex systems also require better heat dissipation and cost-efficient manufacturing methods to stay competitive.
Adopting material solutions that help
New semiconductor designs supporting AI workloads require new classes of materials—ones that can enable high-volume, high-yield manufacturing while ensuring product reliability.
Survey respondents highlighted the following essential materials:
- Encapsulation materials (63.3%)
- Lid/stiffener attach materials (62.5%)
- Underfills between the die and substrate (60.0%)
- Wafer-level handling of large, thin dies (59.2%)
- Thermal management materials (56.7%)
These materials are critical to protecting larger and more delicate chip architectures. They provide mechanical strength to prevent warpage or cracking and are engineered to speed up production with tailored curing and application properties, optimizing factory throughput.
Fueling AI data centers with high-performing, sustainable semiconductors
The semiconductor industry stands at the intersection of performance and sustainability. Enhancing one helps enable the other, creating a synergy that will drive innovation for AI-powered data centers in the years ahead.
At Henkel, we are proud to collaborate with leading semiconductor technologists to develop advanced material solutions that address the twin goals of higher performance and greater sustainability.
Learn more in our full Industry Insights report: “Peer Viewpoint: AI Data Centers are Driving a Semiconductor Revolution.”
