Around tariffs, the AI boom, and rising infrastructure needs: what main differences are you seeing for the semiconductor industry since the last time we spoke?
There’s a lot more attention on semiconductors today, but the fundamentals haven’t changed: make power components that satisfy customer needs. For over 20 years we’ve built power semiconductors, and the surge in EVs, renewables (photovoltaics), and now AI-driven servers—each demanding more power—has simply put our products at the center of more conversations. AI, in particular, is growing rapidly, driving data-center power requirements and, by extension, demand for efficient power semiconductors.
We’re “lucky” to be aligned with these trends, but our approach remains steady: plan carefully and execute for customers. That’s why we built a new fab in China—to grow with our customers as their markets expand. Policy debates, tariffs, and origin rules create noise and complexity, but they don’t change the core imperative: deliver the right component, at the right quality, for the right application.
How has your customer profile been changing, especially as you open the new Nanjing fab?
A few years ago—even two to three—consumer electronics dominated, with a large share going into TV power supplies. Our customers themselves are now shifting: moving from consumer electronics into EV on-board chargers, charging stations, and server power supplies for data centers. We’re growing alongside them, so some changes reflect a market pivot by existing customers.
Consumer electronics remains a piece of the business, but the growth engines are applications that need higher voltages and better conversion efficiency. As customers redirect their product roadmaps toward EV and server power, we align our components and capacity to follow their trajectory—hence the Nanjing expansion.
You manufacture heavily in China. With ongoing tariffs, export controls, and geopolitical uncertainty, how do you balance cost efficiency with supply-chain resilience and compliance?
On the supply side, not much has changed for us because we source locally—whether through partners in India or our own manufacturing—so the production flow is steady. The harder part is the sell-side: navigating where the customer is and how their finished goods move across borders. In many of our product categories, there isn’t a U.S. manufacturer today, so despite tariffs, customers still need these components.
Practically, we see customers paying tariffs—unhappily, but necessarily—because alternatives aren’t in place yet. Free-trade zone dynamics in the U.S. can also mitigate impact when goods enter but ultimately ship back out; the tariff burden is tied to end use in the U.S. Our operating principle is to track where customers’ products will end up and synchronize our logistics and compliance to their strategies.
Given those headwinds, is it still more convenient to keep manufacturing in China rather than, say, the U.S. or elsewhere?
Unfortunately, yes, at this moment. For our components, the U.S. tariff rate can be very high—around 70% and changing often—but the total cost of setting up and running comparable manufacturing in the U.S. would, in our view, exceed that burden. Even at elevated tariff levels, the economics still favor China-based production for now.
We were concerned at first about how much tariffs would suppress demand. In practice, customers have largely absorbed the charges because they need the parts and there isn’t domestic capacity available for these specific components. That could change over the long term, but today the cost structure still points to China.
Has this environment affected your growth? What indicators can you share about the pace since we last spoke—over the past one to two years?
Growth is tied to where the market is moving. We continue to serve legacy consumer electronics, but the newer growth sector is server power—any application involving power conversion benefits from more efficient, higher-voltage components. As voltages rise, conversion losses fall, and efficiency improvesat the system level.
The same logic applies to EVs and photovoltaics: silicon-carbide MOSFETs and diodes support higher voltages and better conversion efficiency, translating to longer driving range between charges, less energy loss when feeding the grid, and lower heat generation in data centers. Less heat means less power spent on cooling—another means for total cost reduction. That’s where we’re focusing on new products.
How are you financing this growth and investments like the Nanjing fab?
New fabs and equipment require substantial capital. We’ve been able to raise funds in the market based on our technology, product fit, and operating history. We are a U.S.-led company—I'm a U.S. citizen—and we have not received Chinese government funding; we’re not a state-owned enterprise.
Our capital has largely come from private investors, often individuals, and from investment arms of large manufacturing companies. We haven’t leaned on big institutional funds; instead, we’ve worked closely with strategic and individual backers who understand the manufacturing landscape.
With AI data centers straining power and efficiency, do you see structural vulnerabilities in the semiconductor supply chain today, or is it fundamentally solid?
There’s more glamour and attention than before, but the core business reality is unchanged: companies exist to serve customers with good products. The industry feels noisier because policy and headlines are constant, yet the success metric remains product quality and fit for the application.
From my decades in semiconductors, I view the current moment as more talked-about, not fundamentally transformed. The path forward is the “old” philosophy: build what customers need, with the right performance and reliability, and you’ll navigate cycles and shifts more effectively than by chasing every headline.
Given that philosophy, how have you consistently delivered products customers want—and how will you keep doing so? What differentiates your approach?
My management style is grounded in engineering: manage by facts and data. No matter the noise, outcomes are determined by measurable performance. We look past narratives to the data—device characteristics, efficiency, reliability, and field results—and use that to make decisions. That’s how we run the company and how we keep improving products.
The other imperative is to constantly ask “What’s next?” New products don’t appear overnight; they require long R&D cycles and capital planning. We’re always defining the next device and the next customer—anticipating higher voltages, better conversion efficiency, and application-specific needs—so we can be here for customers today, next year, and five to ten years out.
