In the broader industry context, this news is hardly a major headline.
It involves a silicon carbide (SiC) design company established in 2025, spun off from its parent company's power device business unit to operate independently in the market; and a packaging and testing manufacturer with over thirty years of experience in discrete devices, whose name exudes the steady aura of an old-established brand. The two parties have signed a strategic cooperation intent for SiC discrete MOSFET packaging.
There was no grand launch event, no investment figures reaching tens of billions, and no viral social media links following the signing. However, I believe it is precisely these seemingly insignificant partnerships that are often closer to the true pulse of industrial evolution than those loud, grand narratives.
This is because they pose a critical question for all SiC practitioners: Who ultimately captures the core value of SiC, at which stage, and through what method? The answer likely lies within this single discrete component that most people would barely glance at.
In recent years, the spotlight on SiC has been almost entirely focused on power modules. Keywords such as main drive inverters, 800V high-voltage platforms, vehicle installation volumes, and On-Board Chargers (OBC) have dominated the vast majority of SiC market narratives, firmly anchoring capital and public attention. It is undeniable—and appropriate—that modules are the protagonists on the central stage.
However, focusing solely on modules causes us to overlook the most solid and extensive other side of SiC: discrete components.
Package types like TO-220, TO-247, TO-252, TO-263, and TOLL—forms that engineers can draw with their eyes closed—are the capillaries through which SiC transitions from automotive mythology to common equipment. Every channel in photovoltaic inverters, every arm in energy storage converters, every module in charging piles, every level in industrial power supplies, and even the increasingly compact fast chargers in mobile phones rely on individual discrete components to concretely realize SiC's performance advantages in terms of efficiency and size.
While modules dance under the spotlight, discrete components silently perform the most frequent, inconspicuous, yet error-intolerant switching operations across countless devices. The long-term imbalance in industrial attention suggests that the value of discrete components has been systematically underestimated.
The maturity of an industry is often not defined by catching up with the most expensive or hottest trends, but by finally being willing to make the fundamentals solid. Discrete components represent the basic skill in SiC that everyone bypasses, yet cannot avoid.
Let us zoom in and first look at the younger party in this collaboration.
Xinzhi Technology was established in 2025 as a wholly-owned subsidiary of Xingan Technology Group, formed by integrating resources from the group's power device business unit for independent operation. It specializes in SiC-based MOSFETs, JFETs, and diodes, as well as Si-based IGBTs and FRDs. Its products cover three voltage platforms: 650V, 1200V, and 1700V, with discrete components spanning mainstream packages from TO-220 to TOLL.
What truly attracts industry attention to this young company is its speed in capturing the market. According to executive interviews, Xinzhi Technology only began consumer electronics market research in March and April 2025. Within just a few months, its 650V SiC MOSFET shipments in the PD fast charging sector reached the ten-million-unit scale.
Beyond enthusiasm, this speed is supported by its technical approach:
Traditional SiC MOSFETs in half-bridge configurations often require negative voltage turn-off to suppress Miller crosstalk, resulting in complex driver solutions and higher peripheral costs. Through high-threshold planar gate design and capacitance matching optimization, Xinzhi has translated zero-voltage turn-off into product language: tolerating turn-off voltages near 0V and driving voltages around 15V, thereby reducing reliance on negative voltage power supplies and active Miller clamping. According to typical tests by the original manufacturer, positive spikes caused by half-bridge crosstalk can be controlled within 1.3V, below the device threshold voltage, leaving engineering margin against false turn-on.
In most consumer-grade fast charging and single-component power supply applications, customers can seamlessly replace existing platforms with SiC discrete components without reconstructing driver circuits. Of course, the trade-off is slightly slower turn-off speed and a minor increase in turn-off losses, but the benefits of system simplification far outweigh these losses.
Landing samples have already emerged, including Silicon Power's SP9477W co-packaged SiC fast charging IC and Somostel's 65W charger, both adopting this zero-voltage turn-off device solution.
A company established for just over a year achieving ten-million-unit shipment volumes in a red ocean market indicates it has at least done one thing right: translating technical language into product definitions that customers are willing to pay for.
Now let us look at the seasoned party in this collaboration.
Good-Ark Semiconductor has deeply cultivated discrete devices for over thirty years, maintaining a complete vertical integration chain from chip design and wafer manufacturing to packaging, testing, and sales. With over fifty series and seven thousand varieties, its rectifier diodes have long ranked among the top domestically. Automotive-grade MOSFETs and IGBTs have completed customer certification and entered mass production, while its packaging and testing lines in Malaysia are steadily releasing capacity.
It is also a dual-core business enterprise combining "Semiconductors + Photovoltaic Silver Paste," with net profit more than doubling year-over-year in the first quarter of this year. In short, this is an old factory that has spent over thirty years honing the craft of turning chips into reliable, mass-producible, and globally deliverable commodities.
The young company runs fast, while the old factory runs steadily. Behind this "steadiness" lies a set of intangibles that cannot be rushed: accumulated failure analysis experience, patience for reliability, every process step on the production line, and craftsmanship repeatedly validated by time. What the SiC industry lacks most today is not people who run fast, but those willing to slow down and do things thoroughly. Good-Ark holds exactly this kind of confidence born from slowness.
With this background, the industrial logic behind the dual-party cooperation becomes clear.
The silicon carbide chain is long, involving substrates, epitaxy, chips, packaging, modules, and applications, with distinct division-of-labor logic at each stage. Design companies are asset-light, fast-moving, and close to the market. However, transforming a chip into a usable, scalable, and automotive-certified device makes the packaging stage unavoidable. Packaging, conversely, is a heavy-asset business competing on yield rates, reliability certification, decades of failure analysis, and process accumulation—it is not something that can be caught up in the short term by simply throwing money at it or hiring staff.
At a forum, I heard an expert remark that left a deep impression: Chips define the upper limit of performance, while packaging determines the realization of that performance. Domestic SiC does not lack good chips; it lacks the last mile of turning good chips into good commodities. By entrusting discrete component packaging to Good-Ark, Xinzhi effectively hands over the initiative in cost, delivery time, yield rate, and reliability on the critical link from chip to device to the party best suited for the task.
This is not merely outsourcing processing to a contract manufacturer; it is binding both parties' technology, capacity, and channels into the same strategic framework. Therefore, it is called strategic layout, not outsourced packaging. For Xinzhi, this means exchanging lower capital investment for a packaging foundation capable of supporting scaled volume growth. For Good-Ark, it means capturing another round of incremental demand from new materials within its familiar power device track.
Mutual benefit and complementary strengths form the fundamental prerequisite for any long-term cooperation.
Looking at the industrial landscape, the market for automotive-grade SiC discrete components is highly concentrated. Industry statistics show that BYD Semiconductor, StarPower, and CR Micro account for nearly 70% of the combined share. BYD Semiconductor alone generates over CNY 1.8 billion in sales, with a market share approaching 40%. For latecomers attempting to confront these giants head-on in major tracks like main drives and automotive standards, success is virtually impossible.
However, the world of SiC extends far beyond main drives. Markets such as PV-storage-charging, high-end power supplies, and consumer power supplies are expanding faster than the automotive sector: charging piles are moving toward higher power, AI server power supplies demand greater efficiency, and even PD fast charging has entered the "multi-port fast charging + full power" era of 3.0, pushing SiC MOSFETs to the center stage. These markets offer wider gaps for latecomers, with more flexible requirements for package forms and costs. Competition among players is not just about grabbing existing shares from others, but crucially about defining increments that no one has yet seen.
SiC is undergoing a fierce price drop. Many view this as a threat, ignoring its other face: Price drops are part of the process of enlarging the cake. Those who can achieve scale during price drops, while simultaneously holding costs and reliability steady, are the ones who truly secure tickets to the next round of competition. Xinzhi has already validated the explosive power of "going from market research to ten-million-unit shipments" in PD fast charging. Now, by solidifying the base of discrete component packaging, it is replicating a proven playbook onto a larger battlefield.
That said, signing a contract does not equal implementation, and strategy does not equal results. The quality of packaging cooperation will ultimately be tested by three hard metrics: Can it truly pass automotive reliability certification? Can yield rates steadily ramp up during mass production? And when price wars deepen, can gross margins be maintained? As SiC substrate and epitaxy costs continue to fall and device prices are repeatedly squeezed, the profit margin for discrete components is far thinner than it appears.
Additionally, there is another point worth considering: The alliance between a design company and a packaging/testing firm is neither adversarial nor unified. Rhythm of R&D iteration, arrangement of capacity priorities, delineation of complaint responsibilities... Any misalignment in these links could reduce a written agreement to mere wall decoration. Binding too tightly may cause the design company to lose flexibility; binding too loosely may fail to achieve strategic impact.
Thus, walking this tightrope beautifully leads to win-win outcomes, while losing balance leads to mutual attrition.
This cooperation is also a microcosm of domestic SiC transitioning from "being able to manufacture chips" to "turning chips into good business."
In recent years, we have become accustomed to measuring progress in domestic SiC by substrate yield rates, chip performance, and module installation volumes. While these are certainly important, they remain within the logic of "manufacturability." The discrete component packaging stage tests a different set of capabilities: cost control, balancing reliability with yield, delivery stability, and patience for certification. These are the keys determining whether a SiC enterprise can convert technological advantages into commercial value.
When unassuming discrete components are installed into photovoltaic inverters, charging piles, server power supplies, and fast chargers—into countless ordinary devices—silicon carbide truly completes its transformation from a "rare laboratory item" to the "tap water of power electronics." By that day, no one will debate whether SiC is usable; they will only scrutinize whether SiC is good enough to use. That will be the true mark of this industry's maturity.
The partnership between Xinzhi and Good-Ark is merely another node in this long industrial chain where variables become constants. China's next breakthrough in SiC may lie hidden within these inconspicuous sub-links, being solidified one by one and lifted up one by one.
Discrete components are small—so small that few people pause to look at them. Yet they are also large—large enough to determine whether an industry can enter millions of households. When the spotlight finally shifts from modules to illuminate these humble packages, it may mark the beginning of domestic SiC growing into a great tree.
Whether this cooperation will materialize remains to be verified by time. But at least, enterprises have begun to take seriously those overlooked capillaries, willing to slow down to connect them. This itself is a signal that the industry is moving forward.