【Expert Interview】Beyond 8 MV/cm: Ga₂O₃ Must Prove Its Value, Not Just Its Performance — Industry Reflections from NextGO Epi’s Dr. Tashun Chou

      The most direct impression from IWGO 2026 was that the gallium oxide community is truly expanding.

      The conference attracted a record‑high number of registered attendees, with 288 participants, marking the largest attendance in the history of IWGO. Compared with the 86 participants at the first IWGO held in Kyoto, the community has grown by more than threefold. This impressive turnout was achieved even with many researchers from mainland China unable to attend due to travel‑related restrictions.

      Comparing with previous IWGO conferences, one trend has become increasingly evident: device-level validation work has surged, with a wide range of device architectures delivering promising high-voltage performance. In particular, vertical devices have essentially emerged as the future direction for gallium oxide power components.

      On the materials side, discussions have become relatively more focused and in-depth. Researchers are shifting their attention toward the origins and impacts of various defects, while exploring novel growth techniques and crystal-plane selections to further reduce defect density and substrate costs. Among these developments, two approaches have attracted considerable attention: Japan’s NCT DG method and Japan’s FOX OCCG method. As breakthrough crucible-free crystal growth technologies, these approaches could, at least in theory, enable the production of 6-inch substrates at costs below $100. However, as with many emerging technologies, mass production still remains far off; as always, a timeline starting from at least five years should be expected.

      Meanwhile, various companies have also showcased their progress in large-size substrates and epitaxial wafers (6-inch scale).

      Overall, gallium oxide research has transitioned from the rapid-growth phase of diverse exploration (2012-2022) into a more stable development stage. Fundamental material studies have largely established a foundation, and the community is converging by consensus around specific technology routes with mass-production potential. For substrates, EFG and VB methods have become the dominant pathways, while epitaxial growth is mainly focused on MOCVD and HVPE technologies.

      Beyond this relatively mature technological framework, remarkable measurement results for Al-alloy-based 2DEG and continued progress in crucible-free crystal growth technologies are adding new possibilities for gallium oxide’s future cost and application scope.

      If I were to summarize the key discussions from IWGO 2026 with two questions, they would be: Which crystal plane is suitable? And when will reliable gallium oxide devices become available?

      To further improve device performance and epitaxial quality, the community has begun expanding its focus beyond the traditionally dominant (001) and (010) crystal planes toward previously less-explored orientations such as (011), (-102), and (100). I welcome this growing interest, as it reflects the continued exploration and maturity of the field. However, I also worry that the discussion may shift away from the most critical priorities.

      For researchers and engineers working on epitaxy and device development, substrate yield and future cost are ultimately far more important than the choice of crystal plane itself. Rather than spending another decade searching for a “perfect” crystal plane — assuming such a plane even exists — it may be more practical to establish a reliable device fabrication process based on a mature, cost-effective substrate orientation.

      At its current stage, gallium oxide urgently needs a reliable device with distinct performance advantages to prove its value to the market.

      During the conference, several semiconductor professionals who had previously worked extensively on SiC pointed out that SiC delivers nearly a hundred-fold improvement over silicon in comprehensive high-voltage performance (breakdown voltage and on-resistance), whereas gallium oxide may offer at most a ten-fold improvement over SiC.

      This raises a fundamental question: Is this performance margin sufficient to justify large-scale investment and sustained industry efforts?

      As the saying goes, “Do not undertake reform unless benefits are hundredfold.” This may be one of the most important questions that gallium oxide practitioners need to address as soon as possible.

      During the roundtable discussion on the conference’s third evening, invited speakers and attendees reached an important consensus on gallium-oxide’s prospects in comparison with SiC: rather than spreading efforts thin, the gallium-oxide community should concentrate on developing reliable devices that can truly deliver its breakdown electric field of 8 MV/cm (or at least approach this value). This is its sole and most dependable competitive advantage.

      As for pricing: gallium oxide still boasts negligible price advantages over silicon carbide. Furthermore, gallium oxide’s genuine cost advantage ought to materialize at the system level, instead of through simple price comparisons of individual die. Drawing on his two-decade-long experience in the SiC industry, Dr. Veliadis from Power America predicts that the first mature gallium-oxide product will likely be a 3.3 kV SBD device, which will serve as the starting point for further advancement toward 20 kV.

      Due to the conference venue, most speakers presented perspectives rooted in European and American viewpoints. Fortunately, Prof. Wen-Hai Huang was among the panelists and brought up “the elephant in the room” — China. Compared with previous editions of IWGO, this year saw participation from representatives of global heavyweights including Mitsubishi and Infineon. This can be partly attributed to the momentum of China, the world’s largest power-semiconductor market.

      China has pulled far ahead of Europe, the United States and other regions in industrialization progress. It holds an order-of-magnitude lead over other countries both in the number of enterprises and supply-chain completeness. For better or worse, this competitive strength remained largely hidden from view because of where the conference was held. Prof. Huang’s remarks reminded the audience that any projection for gallium-oxide’s future would be incomplete without incorporating China as a key variable. In closing, the organizers expressed their hope for joint efforts across the global gallium-oxide community in this direction, looking forward to gallium-oxide delivering remarkable achievements at the next IWGO 2028 in Okayama, Japan.

      If IWGO showcased the possibilities of gallium oxide from a research perspective, then SEMICON Taiwan a few weeks later offered a view of how the same technology is being evaluated in the industrial landscape.

      SEMICON Taiwan has long been a major showcase for the conventional semiconductor industry. Over the past two years, driven by the rapid growth of AI technologies, the scale of exhibitors and attendance reached new highs. A wide range of AI-related technologies took center stage, covering everything from silicon photonics to data center systems.

      With the increasing demand for high-voltage power systems in data centers, gallium oxide, as a candidate material for next-generation power devices, also began to appear at the booths of several material and equipment companies this year, including Japan-based NCT and FOX. Although it remains a supporting role rather than a main focus, gallium oxide is at least no longer a term that requires a lengthy introduction.

      Interestingly, the questions raised on the industrial floor were almost entirely different from those discussed at IWGO. Topics that attracted extensive academic debate—such as crystal orientation selection and defect mechanisms—were rarely brought up here. Instead, industry participants focused on four highly practical questions:

      What is the yield?
      What is the cost?
      When can it enter mass production?
      Who is using it today?

      These four questions happen to represent the areas where gallium oxide currently faces its greatest challenges.

      Such a gap is not surprising—it is a hurdle that every emerging material must overcome when transitioning from the laboratory to manufacturing. However, experiencing both environments firsthand makes the contrast particularly clear: the distance between demonstrating technical potential and achieving industrial adoption remains a critical stage in gallium oxide’s development.

      From IWGO to SEMICON Taiwan, traveling across the two ends of the technology and industry spectrum within just a few weeks led me to a rather straightforward conclusion: the technical potential of gallium oxide has already been demonstrated extensively. What needs to be proven next is its industrial viability—a device with controllable yield, acceptable cost, and clearly defined performance advantages.

      With two years remaining until the next IWGO, which will be held in Okayama, Japan, in 2028, this period may represent a critical window for transforming gallium oxide from a “possibility” into a “practical reality.”

      To date, NextGO Epi has delivered products to dozens of customers mainly in Europe and extending to Asia, including power-device R&D institutes, material-validation laboratories and production-line customers. The product lineup and production capacity keep expanding continuously.