On the eve of Photonics West 2026, we witness the rapid mainstreaming of silicon photonics (and its newer, substrate-agnostic synonym: integrated photonics).
We’ve watched SiPh technology emerge as a tool for bandwidth, efficiency, capacity, and new forms of computing.
But there’s an upshift underway. SiPh needs to scale.
Parallels
This means the industry will undergo the same evolution the semiconductor industry did in the 1980s. Courtesy of the Chip History Center, here is a video showing what a state-of-the-art fab looked like in the early part of that decade. Note the abundance of custom tooling, the manual processing, the human involvement. For that industry to scale, each of those elements had to be supplanted: wafers now are handled and conveyed by robots, humans are largely exiled from the fabs, and a thriving ecosystem of toolmakers emerged, giving us today’s familiar names like Applied Materials, KLA, and ASML.
Today you can provision a fab by writing purchase orders. Back in the day, everything was blank-sheet.
A Language Too: Metrics and Alphabet Soup
Early on, cost-of-ownership became a driving theme in semiconductor manufacturing. Formalized metrics like throughput, yield, cleanliness, and a snowstorm of acronyms like MTBF, MTBA and MTTR all emerged, were codified, were adopted. Fab real-estate became recognized as precious and costly. The vernacular adopted “front-end” and “back-end,” largely distilling down to process-cleanliness requirements. And the 800-pound gorilla of IBM forced the industry to winnow the formats and options down to two.
Fast Forward
For silicon photonics today, similarities abound. The swirl of machinery, process and material options will coalesce around the condensation-nucleus that is TSMC. A tool ecosystem will emerge in navigable form. Its playing-field will be populated by those who learned from semiconductor manufacturing’s example. For them, SiPh’s transition to the mainstream both expands the opportunities and bounds them within norms that will parallel those of the semiconductor industry.
Keyword: Automation
I’ve spent much of my career making markets for automation within photonics manufacturing. With global Internet consumption and now AI driving a three-order-of-magnitude ramp of photonic device production, the time is now, and it is great to see. Just looking back at my own experience: from the earliest single-mode alignment solutions from CSEM via Newport to the first all-digital alignment solutions to the first micro-robotic systems capable of simultaneous alignment optimization across multiple I/O channels, the pace of advancement has been gratifyingly torrid. As it must, to keep pace.
Wait, there’s more
But what of equivalents of those early-semiconductor-industry metrics, that snowstorm of acronyms, those cost-of-ownership considerations?
Towards those, the industry is awakening, and evolving. The ecosystem is emerging: In particular, in the enabling sub-segment of microrobotic aligners capable of fast, multichannel I/O optimization and real-time tracking, the bulky, costly solutions of the past decade are being supplanted by fab-optimized, OEM-friendly solutions:
- Micro-robots requiring two large separate controller boxes are recently joined by units requiring only one… and now, as of Photonics West 2026, none at all.
- With cleanliness taking the spotlight as a vital contributor to yield, micro-robotic implementations with multiple particle-blowing fans are now joined in the marketplace by a novel solution with an integrated, fanless controller that also eliminates multiple sweeping, costly, difficult-to-integrate and lifetime-limiting cables.
- In a market typified by vague or unmentioned reliability specs, integrators, toolmakers, and contract manufacturers are eager for application-optimized variants of proven, robust architectures offering well-established MTBFs exceeding 20,000 hours, with quick in-situ servicing on the rare occasion of an issue, and no fiddly tuning requirements when loads or operating parameters change.
- And with many applications subjected to significant ambient vibration, the robustness and quick settling that comes from stiffness exceeding that of a steel optical post is increasingly valuable.
I don’t want to spoil the surprise, but Photonics West 2026 promises to be a watershed year for the scaling of Photonics. Email me to schedule a personal tour of the available photonic micro-robotic solutions. Can’t make it to San Francisco? Los Angeles in March for OFC will provide another opportunity.
P.S. I’m reminded that my interview on the Manufacturing Matters podcast with Dan McCarthy last year touched on many of these topics. It was a lively discussion! Enjoy.