The legacy of Moore’s Law has shaped generations of engineers and technologies alike. Yet, as device dimensions close in on atomic scales and traditional scaling runs into quantum and thermal walls, the path forward looks vastly different from the past. In a conversation shaped by innovation and shifting benchmarks, Erik Hosler, a technology strategist who highlights emerging architectures beyond conventional silicon, is among those rethinking what semiconductor progress truly means.

The question is no longer about making things smaller. It is about smarter systems, vertical integration, and diversified approaches that extend chip functionality in ways previously unimagined. The industry’s new playbook increasingly includes photonics and MEMS, two fields poised to complement or even replace scaling as we know it. But what role do these technologies really play in keeping Moore’s spirit alive?

Beyond Transistors: Why the Rules Are Changing

For decades, semiconductor advancement was synonymous with shrinking transistors. Smaller features led to faster chips, lower power, and better performance per watt. However, the physics of classical scaling has become stubbornly resistant. We now encounter leakage, variability, and limits of lithographic precision that even EUV cannot overcome indefinitely.

That is why industry leaders are beginning to treat Moore’s Law less as a strict metric and more as a guiding philosophy, one that can be fulfilled by other means. The focus is expanding to include heterogeneous integration, packaging innovation, and entirely new device types. In this broader context, photonics and MEMS are not fringe fields. They are essential components of the next chapter.

MEMS: Tiny Machines, Big Impact

Microelectromechanical Systems (MEMS) are miniature machines that can sense, actuate, or interact with the physical world. They already enable motion sensing in smartphones, pressure sensors in vehicles, and ultrasonic transducers in medical imaging. But their relevance to scaling is even deeper.

As chips integrate more functions into smaller footprints, MEMS offers a way to offload certain tasks such as environmental sensing, RF tuning, and optical modulation without bloating logic circuits. MEMS can also serve as interfaces between digital systems and analog environments, improving responsiveness and reducing computational burden.

In manufacturing, MEMS are already used in metrology tools, enhancing process control at nanoscale dimensions. It improves yield, which in turn drives effective performance improvements even when traditional scaling slows.

The Rise of Integrated Photonics

Photonics, the use of light for data transmission and processing, has been slowly moving from lab benches to foundry floors. Silicon photonics offers a compelling alternative to copper interconnects, which suffer from resistance, capacitance, and thermal limitations as bandwidth demands soar.

In data centers, integrated photonics reduces latency and energy by moving data as light instead of electrons. That is especially valuable in AI applications, where rapid communication between memory and compute units becomes a bottleneck. Beyond communication, photonics shows promise in sensing, quantum computing, and neuromorphic processing.

The real promise of photonics lies in parallelism and speed. Light waves can carry more information more efficiently across a broader spectrum. As fabrication techniques mature, photonic components can be co-packaged or even co-fabricated with CMOS logic, yielding hybrid platforms that redefine performance scaling.

A Hybrid Future: Marrying Electrical, Mechanical, and Optical

The most exciting developments come when MEMS and photonics do not just coexist but work in unison. For example, optical MEMS can tune wavelengths, steer beams, or adjust modulation in real-time, creating adaptive systems for sensing, imaging, and communications.

Hybrid integration platforms combining analog, RF, logic, optical, and mechanical elements are already being prototyped. This approach’s advantage is not just technical but architectural. By distributing workload across different physical domains, such as mechanical motion, light, and electrons, these systems sidestep the limits of any one discipline.

This perspective reframes the scaling problem. Rather than pushing one layer of technology further, we build composite systems where each layer does what it does best. Logic performs compute, photonics manages communication, and MEMS interacts with the world.

Reframing Moore’s Law for System-Level Innovation

The beauty of Moore’s Law was its simplicity, a clean, quantitative benchmark for progress. But the real-world demands of modern computing, from edge AI to biomedical devices, do not fit neatly into transistor counts. These applications need systems that are more adaptable, aware, and efficient, not just faster.

This broader ambition is where MEMS and photonics shine. They allow chips to become more than processors; they become systems that perceive, communicate, and respond. With the right design philosophies, scaling becomes less about lithography and more about integration.

As new use cases continue to emerge, including augmented reality, smart health diagnostics, and autonomous systems, the boundaries between disciplines are blurring. The roadmap ahead is not linear. It is layered.

A Strategic Shift in the Semiconductor Mindset

For decades, success in the semiconductor world meant staying ahead of the node curve. But as we enter the era of 3D stacking, heterogeneous integration, and chiplet ecosystems, innovation means something broader: building smarter systems with diverse toolkits.

This shift is not reactive; it is strategic. Companies are actively investing in MEMS and photonics as scaling enablers. Foundries are refining process nodes that support optical layers and mechanical integration. Design tools are beginning to account for Multiphysics simulation, not just electron flows.

All of this suggests that Moore’s Law is not dead, it’s just being redefined.

This sentiment is echoed in remarks from the SPIE Advanced Lithography symposium, where Erik Hosler notes, “Finally, the solution to keeping Moore’s Law going may entail incorporating photonics, MEMS, and other new technologies into the toolkit.” His statement reflects wider recognition across the industry. The playbook has changed. The tools are broader, the teams more interdisciplinary, and the outcomes more diverse.

Reimagining Progress Through Convergence

The future of scaling lies in composition, not just compression. Photonics and MEMS represent two of the most promising avenues toward that future, not as replacements for silicon but as amplifiers of what integrated systems can achieve.

This development is not about abandoning Moore’s Law but about reinterpreting it through the lens of modern complexity. Progress may no longer be measured in transistor counts but in capability, adaptability, and system intelligence.

By embracing technologies that operate across different physical domains, the semiconductor industry is doing more than finding a way around its challenges. It is writing a new rulebook, one where the laws of light and motion sit alongside those of electrons in shaping the next generation of innovation.