Artificial Intelligence (AI) is advancing at a remarkable pace. It is becoming smarter, faster, and capable of processing massive amounts of data in ways that were unimaginable a few years ago. From powering generative AI applications, enabling autonomous vehicles, to driving medical breakthroughs, every AI system depends on one essential technology: semiconductors. These chips move, store, and process the vast streams of data that AI requires.
However, current manufacturing methods and materials are being pushed to their limits. To meet these demands, the semiconductor industry is undergoing a once-in-a-generation transformation. In the past, major changes in materials or processes often took a decade or more to fully implement across the industry. But AI’s explosive growth has compressed those timelines dramatically. What used to be a gradual evolution is now an accelerated revolution.
Evolving for the AI Era
Building advanced chips such as NAND, DRAM, and logic devices, requires extreme precision, complex engineering, and materials that perform reliably at nanoscale dimensions. For rapidly growing AI workloads, the predictable paths of scaling by packing more transistors into smaller spaces are no longer enough.
Traditional processes face physical and technical limits. As chips complexity increases, electrical signals must travel through increasingly narrow and intricate connections. These connections can become bottlenecks, slowing down processing speed and increasing the risk of electrical shorts. To overcome these hurdles, companies like Lam Research are rethinking the fundamentals of semiconductor manufacturing.
A New Approach to Metallization
One of the most important changes is in the metals used to connect circuits inside chips. These connections, known as interconnects, are vital for transmitting electrical signals quickly and efficiently. For more than 25 years, tungsten has been the metal of choice. However, tungsten requires the addition of a barrier layer, typically titanium nitride (TiN), to improve adhesion and prevent electrical leakage. The problem is that this barrier layer has a much higher resistivity than pure tungsten, which reduces overall performance.
Enter molybdenum, a metal that offers several advantages for next-generation chips. Unlike tungsten, molybdenum doesn’t require a highly resistive barrier layer. This means more space is available for the conductive metal itself, lowering electrical resistance and allowing signals to move faster. In fact, studies show that molybdenum-filled contacts can have more than 50% lower resistance compared to those using the traditional tungsten-and-barrier approach.
Molybdenum can also simplify manufacturing by reducing the number of process steps, improving efficiency, and supporting the continued scaling of chip designs. While tungsten will still be used in certain applications, the shift to molybdenum represents a generational leap in semiconductor metallization; one that will help power the AI era and the future of high-performance computing.
Breakthroughs in Etch and Deposition Technologies
While the change in metals is significant, it is only part of the transformation. Equally important are innovations in etching and deposition, the processes used to shape and build the microscopic structures within chips.
One recent development involves the use of cryogenic processing environments to achieve deeper and more uniform etch profiles in memory devices. Operating at extremely low temperatures allows plasma-based systems to maintain tighter control over sidewall shape and depth consistency, even when forming channels that extend several microns below the surface. This approach can improve both structural accuracy and throughput, making it better suited for high-volume manufacturing.
On the deposition side, atomic layer deposition continues to gain importance as manufacturers seek precise control over material thickness at the atomic scale. Newer ALD platforms have been engineered to handle emerging metals such as molybdenum while maintaining uniform coverage across increasingly complex geometries. Bringing these capabilities from research settings into large-scale production has required extensive process refinement, but it represents a key step toward integrating next-generation materials into commercial devices.
Building the Foundation for the AI Era
The changes happening in semiconductor manufacturing are not just about meeting today’s needs but also about preparing for the future. These advancements will enable faster data processing, more efficient energy use, and greater overall performance in computing devices. Just as important, they will ensure that the industry can continue to scale its technologies in line with the increasing demands of AI and other emerging fields.
In essence, the semiconductor industry is laying down a new technological foundation, that will support innovation for decades to come. This is a generational shift, and it is happening in real time. The move from tungsten to molybdenum, combined with breakthroughs in etch and deposition, marks a turning point in how chips are made.