For more than five decades, silicon has been the bedrock of semiconductor advancement. It powered Moore’s Law, allowing transistors to shrink while delivering ever greater performance at lower cost. But today, as scaling becomes more difficult, silicon alone will not be enough. Erik Hosler, a semiconductor systems consultant with deep knowledge of lithographic scaling and cross-domain integration, recognizes that the future lies in expanding the toolkit beyond traditional materials.
This expansion reflects a shift from reliance on silicon-based shrinking to a new model based on performance, functionality, and diversity of technologies. As we move into a post-silicon mindset, it is no longer about doing more with less. It is about doing more with more kinds of materials, structures, and platforms. Photonics, MEMS, and emerging substrates are taking center stage, each playing a role in carrying forward the spirit of Moore’s Law.
The Limits of Traditional Scaling
Shrinking transistors is no longer a simple or cost-effective strategy. At nodes below 5 nanometers, quantum effects disrupt switching behavior, leakage currents increase, and process variability undermines yield. Every step forward demands enormous engineering effort and capital investment.
Yet, the appetite for progress has not diminished. Consumers expect devices that are faster, smarter, and more energy efficient. Artificial intelligence workloads, edge computing, and real-time connectivity demand advances that traditional silicon scaling can no longer deliver. That is why the industry is looking outward toward new materials, new architectures, and new domains.
Silicon Is Still Important, But No Longer Alone
It is worth noting that silicon will continue to be essential for the future. It remains the most well-understood and scalable semiconductor material, but it now shares the stage with others.
Platforms such as Gallium Nitride (GaN) and Silicon Carbide (SiC) are growing in popularity for high-power and high-frequency applications. Two-dimensional materials like graphene and transition metal dichalcogenides promise novel approaches to conductivity and miniaturization. Hybrid systems increasingly pair silicon logic with photonic interconnects or MEMS-based sensors to broaden functionality. The torch of progress is not being handed down. It is being passed among multiple runners, each suited to a different part of the course.
Photonics for High-Speed Communication
One of the clearest areas where silicon shows limitations is communication. As data volumes rise, electrical interconnects face constraints related to resistance, heat, and signal loss. Photonics sidesteps these limitations by transmitting information via light.
Photonics enables higher bandwidth at lower power over greater distances, making it essential for next-generation data centers, AI clusters, and chiplet architectures. Silicon photonics, the integration of photonic components into CMOS-compatible substrates, is already seeing widespread adoption.
But this is only the beginning. As computing becomes more distributed and communication more essential, optical solutions will move deeper into device architecture, becoming as foundational as transistors themselves.
MEMS as Sensory Enhancers
Microelectromechanical Systems (MEMS) are another platform shaping the future. These miniature mechanical devices, often built on silicon, allow chips to sense motion, pressure, sound, and environmental conditions.
In mobile devices, MEMS provides critical input for orientation, navigation, and interaction. In automotive, power safety systems and environmental monitoring are important. In medical and industrial contexts, they enable precision sensing in compact, energy-efficient formats.
As chips move from processors to interactive agents, MEMS devices allow silicon to interface more meaningfully with the real world. Their integration into scaling roadmaps ensures that capability scales alongside computation.
A Broader Perspective from the SPIE Symposium
This multifaceted approach to innovation was a key theme at the SPIE Advanced Lithography symposium. Industry leaders shared insights across lithography, materials, design, and integration, emphasizing that no single breakthrough would suffice.
Instead, they pointed to the value of emerging platforms that could extend the scaling narrative beyond feature size. Erik Hosler shares, “Finally, the solution to keeping Moore’s Law going may entail incorporating photonics, MEMS, and other new technologies into the toolkit.” This sentiment captures the growing consensus that Moore’s Law continues, but its expression has changed. It now resides in the performance gains that come from many sources working together.
The Role of Integration Technologies
If diverse technologies are to succeed, they must be effectively combined. Ts where advanced packaging and heterogeneous integration come into play. By assembling systems from multiple dies and technologies, engineers can optimize each function without relying on a single process node.
A silicon chiplet for logic might sit alongside a photonic die for communication and a MEMS sensor for environmental input, all within a single package. This model decouples performance from raw transistor density, allowing systems to scale even when traditional lithography stalls. Integration has become the new frontier of innovation.
Design Tools Must Keep Up
To support these heterogeneous platforms, design automation tools must be developed. Designers need simulation environments that account for thermal effects in photonics, mechanical behavior in MEMS, and process variations in advanced materials.
Cross-domain verification is no longer required. It is essential to ensure that innovative technology combinations behave as expected in real-world conditions. As the platforms expand, so must the digital tools that orchestrate them.
Implications for AI and Edge Computing
The emergence of new platforms directly benefits high-demand applications like AI and edge computing. These workloads require real-time response, high data throughput, and low energy consumption within constrained environments.
Photonics supports fast, low-latency communication across chiplets or devices. MEMS enables context awareness and adaptive behavior. Novel materials enhance energy efficiency and temperature tolerance. Together, they deliver the kind of system-level performance that Moore’s Law always promises, even if it no longer arrives through simple miniaturization.
Education and Infrastructure Must Follow
With more platforms in play, engineers must be trained in broader skill sets. Universities are beginning to offer interdisciplinary programs that include optics, materials science, mechanical systems, and integration.
Fabrication infrastructure must also be adapted. Foundries and packaging facilities are investing in process capabilities for GaN, SiC, photonics, and MEMS. Standardization efforts aim to make heterogeneous integration more scalable and predictable.
The industry’s ability to scale now depends on its ability to teach and build diversity.
A Torch Passed, Not Dropped
The legacy of Moore’s Law is not confined to silicon. It reflects a mindset that performance must keep advancing, even as the tools and materials evolve. That momentum now comes from a combination of platforms, each contributing to the broader system.
Photonics improves speed, MEMS enables sensing and interaction, and new materials open paths to greater efficiency and adaptability. These technologies are shaping a post-silicon era that stays true to Moore’s original intent by pushing meaningful progress where it matters most.
The torch is still burning. It now illuminates a wider path forward.
