You have been deeply involved in the semiconductor industry for over three decades. How would you describe the sector’s evolution, and what stands out most compared to your early days of research?
The semiconductor industry has always been rooted in collaboration—it draws on the best skills and capacities from around the world. Over the decades, we have seen titans rise and fade. In the 1990s, companies like NEC, Motorola, and Digital Equipment Corporation were all on the cutting edge. Today, none of them are involved in semiconductors.
There’s a constant passing of the torch. Now, with what’s happening in AI, leaders are changing again. Government involvement has also intensified as semiconductors underpin everything from national security to AI, and the way in which data will be created and processed in the future. History rhymes—leaders evolve, but the story stays the same.
Given today’s intense competition, what do you think sets future leaders apart?
Generating knowledge from data and learning from your environment are now core capabilities. But more than collecting data, it's about connecting your limited data drops to the ocean of global insights—and doing so collaboratively.
Leadership demands close coordination across the supply chain. Whether you're building memory chips or analog circuits, working hand-in-hand with your ecosystem is crucial—suppliers help you understand tools; customers help you understand tolerance requirements. It’s a team sport: one where the stakes are in the tens of billions.
People often underestimate how sophisticated semiconductor manufacturing is. The gap between being top-notch and slightly off can be ruinous. Just being second or third to market can be the difference between profit and loss. When you're making $5 billion R&D bets and building $15–20 billion factories, you can't afford operational inefficiencies.
Could you explain in simple terms how PDF Solutions contributes to this “team sport”?
Manufacturing a chip takes about three months and involves over 1,000 steps. If something goes wrong early, you might not catch it until the end—risking hundreds of millions of dollars. Our software helps detect issues early by collecting and interpreting equipment data in real time, flagging unusual patterns before they cause widespread waste.
We also connect manufacturers with equipment suppliers. If a machine behaves unexpectedly, our secureWISE software allows engineers across continents—say, one in Arizona and another in the Netherlands—to collaborate instantly. Together, they can troubleshoot issues before the problem spreads. One platform identifies anomalies, the other enables real-time collaboration—preventing months of loss in just one call.
You recently launched the Sapience Manufacturing Hub Enterprise. What kind of challenges can this solve for your customers?
Sapience was built to remove human bottlenecks from AI-enabled manufacturing. Decisions like where to send a chip next—based on its test data—shouldn't require human intervention. If they do, production timelines can stretch from three to five months due to engineering holds.
Sapience connects engineering with enterprise resource planning and manufacturing execution systems, allowing AI to orchestrate these workflows without manual input. This low-code integration layer enables autonomous decisions, such as routing chips based on demand, factory readiness, or end-market requirements.
Engineering and manufacturing form the backbone of semiconductor company performance, yet critical business decisions often occur in isolation from the rich operational data these functions generate. This disconnect creates blind spots that can undermine strategic planning and operational efficiency.
Sapience addresses this fundamental challenge by integrating granular engineering and manufacturing insights directly into enterprise-level decision-making processes. By providing executives and business leaders with unprecedented visibility into production metrics, yield data, and process performance, Sapience transforms how semiconductor companies align their strategic initiatives with operational realities.
This data-driven approach enables more informed resource allocation, faster response to manufacturing challenges, and better alignment between business objectives and production capabilities—ultimately driving improved performance across the entire organization.
PDF Solutions reported 16% year-over-year revenue growth and record Sapience bookings in Q1 2025. To what do you attribute this momentum?
It’s always a marriage of strong products and market timing. You could build the best horse and buggy, but in the 21st century, it won’t sell. Today, three industry trends are propelling growth. First, the move to 3D chip packaging for higher density and faster signal transfer—requiring new tools to predict and solve yield issues.
Second, geopolitical shifts are decentralising manufacturing via the US, Euro, Japan, and India Chips Acts—making remote collaboration crucial. Third, with talent shortages, AI must augment human decision-making. Our software sits at the intersection of these trends—built on a common platform and database to help customers advance, collaborate, and scale AI without friction.
The secureWise acquisition was significant—$130 million, equivalent to 15% of your company’s market cap. What made it so valuable to your strategy?
secureWise began at IBM and now links 200 fabs globally to their equipment suppliers through a secure, private network. These machines are incredibly expensive (up to $300 million each) and the data they generate is enormous. secureWise enables real-time, remote collaboration to ensure they run optimally.
It also offers a safe method for deploying software or AI updates without risk of viruses—a real threat, as one factory-wide outage can halt operations. We saw secureWise as essential for 21st-century manufacturing. Combined with our own platforms, it’s like peanut butter and chocolate—better together than alone. We saved a decade of development and gained trust from secureWise’s 20-year track record.
You have also stepped into an educator role—launching blogs, events, and university courses. What inspired this educational push?
We spun out of Carnegie Mellon University and used to feel a bit nerdy about our academic roots. But five years ago, we embraced it. The industry was aging, and fewer young people were entering. I remember a conference where I asked: “How many of you would recommend the chip industry to your kids?” No one raised their hand. That’s when I knew Moore’s Law was dead—not for technical reasons, but due to waning interest.
To be vibrant, our industry needs fresh thinkers who also understand foundational knowledge. You can’t innovate without first learning the classics, like Bach before composing your own. So, we partnered with Carnegie Mellon and Intel to teach AI in manufacturing, offered workshops, and started publishing blogs—not to preach, but to give new entrants a base to build upon.
Finally, what’s next? What are your top priorities for the next year or two?
Everything takes years to realise. Sapience and our model ops platform were built as foundations. The paradigm we’re now trying to shift is that volume equals dominance in semiconductor manufacturing. As AI drives demand for the most advanced nodes without necessarily high volumes, the industry must shift toward learning and ramping effectively on smaller data sets.
Our goal is to enable effective manufacturing for these processors, even in locations with no existing ecosystem, like Dresden or Texas. That’s the future we’re building for.
