On August 13, 2026, Hangzhou Garen Semiconductor announced the closing of its hundreds-of-millions-yuan Series A funding round; four days later, its local peer Fujia Gallium also officially announced the completion of a new round of financing exceeding 100 million yuan. Two companies, two technology routes, and two rounds of intensive capital injection occurred in quick succession in the city of Hangzhou. The list of investors includes Shenzhen Capital Group, CRRC (China Railway Rolling Stock Corporation), Quzhou Dongfeng Group, Fangguang Capital, Shanghai Science and Technology Innovation... National-level capital, industry leaders, local state-owned capital, and market-oriented VCs are densely intertwined. I believe this is no coincidence, but rather an industrial manifesto written in hard cash: Gallium oxide, a material referred to by the academic community as the 'fourth-generation semiconductor,' is rushing from the laboratory to the production line at an unprecedented speed. Looking at this timing makes it even more interesting. The outline of the '15th Five-Year Plan' has just explicitly included 'promoting the industrialization and development of ultra-wide bandgap semiconductors such as gallium oxide and diamond,' marking the first time gallium oxide has been incorporated into a national five-year plan. Once the policy floodgates opened, a torrent of capital rushed in. Fujia Gallium and Garen Semiconductor happen to be standing at the forefront of this wave.
01. Two Strategies, One Goal If the gallium oxide industry is compared to a race, Fujia Gallium and Garen Semiconductor are the two runners running side by side in the leading group. They have different paces, but the same direction. Fujia Gallium is a 'full-chain heavy artillery' player. Founded at the end of 2019, relying on the technological foundation of the Shanghai Institute of Optics and Fine Mechanics (SIOM), Chinese Academy of Sciences (CAS), it follows the Vertical Bridgman (VB) method route. It took the lead internationally in introducing AI technology into the crystal growth process, achieving 'one-click crystal growth' and obtaining international patent authorization. In March 2026, this company did something that made global peers take notice: it used the VB method to prepare the world's first 12-inch gallium oxide single crystal, with a diameter of 305 mm, breaking the global record. From 6 inches to 12 inches, the theoretical output of single-chip devices is 4 times that of 6-inch and 2.25 times that of 8-inch. This is not a simple superposition of size, but directly smashes the cost ceiling of gallium oxide devices in half. *Fujia Gallium's 12-inch gallium oxide single crystal (diameter 305mm) More critically, Fujia Gallium has built a fully independent closed loop of 'core equipment—crystal growth—substrate processing—thin-film epitaxy.' It is not only the only company in the world with 12-inch gallium oxide single crystal preparation technology, but its epitaxial thin-film mobility has also reached the highest value reported internationally, approaching the theoretical limit of gallium oxide materials. In June 2026, the company signed a strategic cooperation agreement with device-side leader Marching Power, officially moving from 'material breakthrough' to 'full-chain synergy.' At the capital level, Quzhou Dongfeng Group (601515), as the second largest shareholder, holds 16.77% of the shares. Wuchan Zhongda indirectly participates through an industrial fund. The dual endorsement of the national team and listed companies makes Fujia Gallium's industrial position more solid. *In July 2026, Fujia Gallium successfully transferred the 8-inch gallium oxide MOCVD epitaxial process to 12 inches for the first time. The picture shows a comparison of the epitaxial thickness distribution of 8-inch (left) and 12-inch (right) gallium oxide.
Garen Semiconductor, on the other hand, is a 'technology route innovation' player. Founded only in September 2022, its founder Zhang Hui is a professor of materials science at Zhejiang University, and its chief consultant is Yang Deren, an academician of the CAS—a towering figure in the domestic silicon materials field. The team's background determines its differentiated approach: while others use the Edge-defined Film-fed Growth (EFG) method, it uses its self-developed 'casting method'; while others are choked by the high cost of the precious metal iridium, it cuts iridium usage by 80%; while others are still worrying about the yield of 4-inch wafers, it is already delivering 6-inch homoepitaxial wafers in batches. The essence of the casting method is to achieve directional solidification of the melt in a controllable thermal field, which is consistent with the underlying logic of silicon single crystal growth. It is precisely for this reason that a team with a silicon materials background can achieve an 'overwhelming disruptive advantage' in the field of gallium oxide. Garen Semiconductor has reduced the cost of a single substrate wafer by more than 80% and increased the wafer output to 3 to 4 times the original amount. The multiplicative effect of these two figures is expected to push the cost of 6-inch substrates below 2,000 yuan, which will significantly reduce downstream device costs. *Garen's 6-inch gallium oxide homoepitaxial wafers have achieved batch supply.
In July 2026, the company built the world's first mass production line for gallium oxide homoepitaxy compatible with 6/8-inch wafers, becoming the world's first and only supplier of commercial 6-inch gallium oxide homoepitaxial wafers. The products are not only delivered to leading domestic wafer fabs, but also stably supplied in batches to the German technology company NextGO.Epi, taking the first step for Chinese gallium oxide materials to go global.
02. What Are Investors Really Betting On? To understand why capital is concentrating its bets on gallium oxide in 2026, we must return to a fundamental question: Why is gallium oxide more worthy of betting on than silicon carbide (SiC)? The answer is actually quite simple. The bandgap of gallium oxide is about 4.8-4.9 eV, while silicon carbide is only 3.2 eV; the critical breakdown electric field of gallium oxide is about 8 MV/cm, while silicon carbide is only about 3 MV/cm; the BFOM (Baliga's Figure of Merit, the core indicator for measuring the cost-effectiveness of power devices) of gallium oxide is as high as about 3444, which is more than ten times that of silicon carbide. Translated into plain language: for devices with the same voltage rating, gallium oxide can be made thinner, smaller, and with lower losses. In ultra-high voltage scenarios above 8kV, such as ultra-high voltage power transmission, rail traction, and megawatt-level data center power supplies, silicon carbide is already approaching the dual ceiling of physics and cost, while gallium oxide is just opening the stage. But what really excites investors is not the performance parameters, but the cost structure. Silicon carbide is grown using the vapor phase method, which is slow, limited in size, and requires expensive equipment; the cost of a single 6-inch substrate reaches thousands of yuan. Gallium oxide is the only wide bandgap material that can be grown at atmospheric pressure using the melt method just like silicon. Its growth rate is more than 10 times that of silicon carbide, and the theoretical substrate cost is expected to drop to one-third or even lower than that of silicon carbide. When 'wider bandgap' and 'lower cost' appear simultaneously, gallium oxide is no longer a high-end niche supplement, but has the qualification to reshape the entire power semiconductor supply chain. This is exactly the 'option logic' in the eyes of investors. The equipment investment for a single silicon carbide production line is often hundreds of millions of yuan, and it is incompatible with gallium oxide processes, making the switching cost of material systems extremely high. Against the backdrop of already overdrawn valuations in the main silicon carbide track, capital spilling over into gallium oxide is essentially laying out in advance for 'what might come after silicon carbide.' Fangguang Capital and Shenzhen Capital Group co-led the investment in Garen Semiconductor, Yuanchuang Fund bet on Fujia Gallium, and CRRC entered the market directly as a downstream application party. Every single penny is saying the same thing: the arrival of gallium oxide is not a question of 'whether it will come,' but 'when it will come.' The density of financing itself also illustrates the point. Since 2026, domestic ultra-wide bandgap semiconductor financing has been exceptionally dense: two rounds of hundred-million-yuan financing for Fujia Gallium, a Series A of hundreds of millions of yuan for Garen Semiconductor, an A++ round of 110 million yuan plus a Pre-B round of over 150 million yuan for MIG Semiconductor (totaling nearly 500 million yuan across six rounds), and consecutive landings of angel rounds and strategic investments for GaChuang Future. According to industry data, the total domestic investment in ultra-wide bandgap semiconductors is expected to exceed 21 billion yuan in 2026, doubling that of 2023. This density is not a bubble, but a real signal of the industry switching from the 'technology R&D cycle' to the 'order validation cycle.'
03. Reflections Behind the Boom However, the enthusiasm of capital cannot conceal the harsh reality of the industry. Three major mountains lie ahead for gallium oxide, which all gallium oxide companies cannot bypass. The most immediate challenge is heat dissipation. The thermal conductivity of gallium oxide is only about 10-27 W/m·K, which is about one-thirteenth that of silicon carbide and one-fifth that of silicon. When operating under high voltage and high current, heat inside the device accumulates rapidly but cannot be dissipated, causing the carrier mobility to drop rapidly, and in severe cases, leading directly to thermal breakdown. In 2025, the team of Academician Hao Yue from Xidian University published an ingenious solution in Nature Communications: introducing graphene as an interface buffer layer between gallium oxide and diamond, reducing the interface thermal resistance to one-tenth of traditional technologies. But this is still a laboratory breakthrough, and there is still a distance to mass production. Backside thinning, diamond bonding, Ga₂O₃-on-Diamond heterogeneous integration... all heat dissipation solutions are still on the way from 'proof of concept' to 'engineering.' The second challenge is p-type doping. The basic principle of semiconductor physics is that p-type and n-type must be complementary to create efficient bipolar devices and CMOS circuits. However, the valence band maximum of gallium oxide is composed of O 2p orbitals, the hole effective mass is large, and the acceptor energy level is deep. To date, effective and repeatable p-type conductivity has not been achieved. This directly limits the design freedom of bipolar devices and CMOS-based architectures. Consequently, gallium oxide power devices as a whole remain in the unipolar camp—SBDs were the first to be commercialized in small batches, lateral MOSFETs are close to commercialization (Jiufengshan Laboratory made a 9.02kV lateral MOSFET based on domestic homoepitaxial wafers in 2026), and vertical UMOSFETs/JFETs are rapidly climbing in the kilovolt to 9-kilovolt range. To bypass the deadlock of p-type doping, the domestic academic community has taken two parallel routes: heterojunction integration (p-NiO, etc.) + surface electric field engineering (field plate/dielectric interface modulation). The 9kV device from Jiufengshan belongs to the latter; however, neither route has yet converged into an automotive-grade, reproducible, and standardized device system. For this view, please refer to my previous article, 'After the 9000V Breakthrough, Let's Not Repeat the Old Path of Silicon Carbide!' The last challenge is the lack of an industrial chain ecosystem. Silicon carbide has undergone nearly 20 years of ecosystem construction, possessing a complete supply chain from substrates, epitaxy, device design to module packaging, as well as an automotive-grade validation system. Domestic gallium oxide has crossed the 'sample delivery' stage, and 6-inch substrates and homoepitaxial wafers have entered the initial stage of small-batch mass production and commercial delivery. However, from wafer yield and device reliability databases to automotive-grade certification, there is still a significant gap compared to the mature ecosystem accumulated by the silicon carbide industry. The strategic cooperation between Fujia Gallium and Marching Power, the joint research between Garen Semiconductor and Sanan Optoelectronics, and the batch supply to CR Micro are essentially all about making up for lost time—making up the lesson of full-chain synergy from 'substrate epitaxy—chip devices—terminal applications.'
04. Global Competition Expanding the vision globally, the competitive landscape of gallium oxide becomes clearer. Japan is a traditional powerhouse. Novel Crystal Technology (NCT) has long supplied nearly 100% of the world's gallium oxide substrates and holds about 60% of core invention patents. In March 2026, NCT began batch delivery of 150mm (6-inch) β-Ga₂O₃ substrate samples, and plans to launch epitaxial wafers in 2027, achieve full mass production in 2029, and supply 200mm (8-inch) substrates in 2035. The EFG method it uses has mature technology, but relies heavily on the precious metal iridium, keeping costs high. To this end, NCT is developing the 'DG method' that does not require iridium crucibles, attempting to achieve cost disruption at the mass production stage. The United States regards gallium oxide as the core material for next-generation military and civilian power electronic systems. The Department of Defense has invested up to $280 million in the R&D of gallium oxide power switches, and DARPA has launched a special project focusing on device-level application development. European research forces are dispersed, but they have outstanding strength in basic research such as material characterization and device physical models. The Fraunhofer Institute in Germany and the EPSRC in the UK both have layouts in this area. In this global race, the strategy of Chinese companies is very clear: using the combination punch of large size + low cost to quickly erode Japan's first-mover technological advantage.
- (a) Cross-sectional schematic and key device manufacturing details of the junction termination extension NiO/β-Ga2O3 heterojunction diode, (b) performance comparison with reported gallium oxide Schottky diodes and heterojunction diodes.
Fujia Gallium's 12-inch single crystal expands the size lead to at least one generation; Garen Semiconductor's casting method tears open the gap for industrialization from the cost end; MIG Semiconductor positions itself in the two major markets of optical communication and power through its dual layout of indium phosphide + gallium oxide. More importantly, China has the world's largest semiconductor consumer market. In 2026, the demand for gallium oxide-related products accounted for 31% of the global total, with the power electronics field contributing over 58%. The vast downstream application scenarios are a natural advantage that no overseas competitor possesses.
- (a) Schematic diagram and process flow chart of the heterogeneous PN gallium oxide junction field-effect transistor, (b) transfer characteristics under different drain biases, (c) output characteristic curves, and (d) breakdown characteristic curves.
But objectively speaking, Japan still has years of accumulated advantages in long-term reliability testing of automotive-grade devices, epitaxial process details, and patent barriers. Once NCT's DG method is successfully mass-produced in 2029, it will directly impact the cost advantage of Chinese companies. Pilot production lines of European and American research institutions are also being accelerated. The 'three-year lead' of Chinese companies is not a cause for complacency, but a fragile advantage that needs to be continuously consolidated through continuous R&D investment.
05. Policy Catalyst and Industrial Inflection Point 2026 is a 'major policy year' for China's gallium oxide industry. The industrialization of gallium oxide is written into the '15th Five-Year Plan' for the first time. The Ministry of Science and Technology has included it in the key R&D plan, supported by a 2.6 billion yuan industrialization guidance fund. At the local level, Zhejiang has become the only province in the country to include gallium oxide and diamond in its provincial '15th Five-Year Plan' for new industrialization, aiming for the integrated circuit industry revenue to reach 450 billion yuan by 2030. Shanghai Lingang has released the country's first local special policy for the fourth-generation semiconductor, aiming to gather no less than 50 enterprises and break through an industrial scale of 5 billion yuan by 2028. Provinces and cities such as Hubei, Anhui, Guangdong, and Sichuan have followed up intensively, providing tens of millions of yuan in support across eight major areas, including major project investment, chip tape-out, and testing validation. But if we think from another perspective, the density of policies reflects the strategic anxiety at the top level. The core raw material of gallium oxide is metallic gallium. Since 2023, China has implemented export controls on gallium-related items, and reiterated multiple times in 2025 that it applies to gallium oxide, highlighting its strategic resource attributes. On a global scale, whoever achieves the industrialization of gallium oxide first will master the pricing power of the next generation of power electronics. This is destined to be a smokeless material battle.
06. The Real Validation Checkpoint Returning to Fujia Gallium and Garen Semiconductor themselves. The two companies are not in zero-sum competition, but together constitute the 'dual engines' of China's gallium oxide industry. Relying on the 12-inch global record and full-chain system capabilities, Fujia Gallium has established a leading position in cutting-edge technical indicators and enterprise scale; relying on the cost disruption of the casting method and batch supply in the international market, Garen Semiconductor has taken the lead in the speed of commercialization. I think their competition is more like a 'friendly race,' competing on who can push the yield past 80% first, compress the cost to the critical point where customers are willing to place orders, and get the devices into the validation process of automakers. For investors, gallium oxide currently looks more like an 'option for three years from now' rather than 'current cash flow.' The real catalytic nodes will appear in 2027-2028: when the yield of 6/8-inch substrates stably breaks through 80%, when gallium oxide power devices complete 10,000-hour reliability validation on the NEV (New Energy Vehicle) or power grid side, and when leading device manufacturers begin to purchase domestic epitaxial wafers in batches. Market institutions generally predict that 2026-2027 will mainly focus on sampling and small-batch trials by automakers and photovoltaic companies, and the scale installation wave will not come until after 2028. Fuji Keizai calculates that the global market size for gallium oxide power devices will be about 154.2 billion yen (about $1.22 billion) in 2030; the more optimistic Stratistics MRC expects that the overall global gallium oxide market will reach $161.28 billion by 2034, with a CAGR of 15.7%. The predicted figures from the two institutions differ greatly, but the direction is consistent: gallium oxide is an acceleratingly expanding market, it just requires patience for the realization pace.
07. The Ultimate Winner Qi Hongji, founder of Fujia, said in an interview: 'Gallium oxide is at a critical inflection point for the full-chain integration from materials to devices.' In less than four years, Zhang Hui has pushed Garen Semiconductor from a laboratory to the doorstep of the world's first 6/8-inch mass production line. I think this inflection point means two things: the direction is already clear, and the most difficult path is still under our feet. For investors, this is an option that requires patience to wait for realization; for industry practitioners, this is a marathon about yield, cost, and reliability; for China, this is a historic opportunity to shift from 'following' to 'changing lanes and leading' in the field of semiconductor materials. The patience left by the market for everyone is only about three to five years. Whoever can take the lead in running through the full-chain closed loop of 'substrate—epitaxy—device—module' within this window period will carve their name into the map of the next generation of power electronics.