In 2011, in an unassuming factory building in Haidian, Beijing, a few individuals were wrestling with several pieces of equipment.
At that time, silicon carbide (SiC) was still an extremely obscure term in China. Investors would politely nod after hearing the pitch, calling it "a story for ten years later," and then turn their attention to chasing the booms in photovoltaics and LEDs. Customers would inspect samples repeatedly, with their first question always being, "Can this be cheaper?" Even suppliers advised them not to burn cash, saying that even Europe and the US hadn't figured it out yet, so why the rush?
They didn't argue; they just moved equipment into the cleanroom one by one and ran processes over and over again. At that time, no one believed this could succeed, including perhaps themselves. But they did it anyway: transforming SiC devices from lab samples into reliable components capable of stable shipment, automotive-grade certification, and integration into complete vehicles and charging piles.
This path has been walked for fifteen years.
Fifteen years later, today, in Shunyi, Beijing, an 8-inch SiC production line facility has topped out. This silent gray structure is the key chapter added to Global Power Technology' fifteen-year IDM narrative.
From 4-inch to 6-inch, and then to 8-inch wafers, Global Power Technology completed the leap across three generations of wafer sizes in fifteen years.
And this leap was never a smooth ride.
In April 2011, Chen Tong founded Global Power Technology in Beijing.
Across the ocean, Cree had not yet renamed itself Wolfspeed; SiC MOSFETs had just moved from labs to commercial catalogs. Infineon's SiC was still stuck in the diode era, with MOSFETs only peeking through the door. ROHM had just mass-produced planar SiC MOSFETs in 2010, while trench-gate structures were only pushed toward scale production in 2015. What about domestic players? It was virtually a blank slate.
Chen Tong once used an analogy: chip manufacturing is like building a house inside a hair strand, stacking layer by layer upwards, at micron and nanometer scales, without a ruler. Later, media used this phrase to describe Global Power Technology' early team as "people building houses inside hair strands." This metaphor is not exaggerated; the process precision of SiC devices indeed requires multi-layer structures within cross-sections as thin as a hair.
In 2011, annual sales of new energy vehicles (NEVs) in China were less than 10,000 units, still in the stage of buses and demonstration operations. The photovoltaic industry was amidst the brutal storm of its first round of reshuffling. Charging piles, energy storage, and data center power supplies—the main battlefields for SiC today—had no shadow yet.
Chen Tong chose to go All-in on SiC devices at this point not because he saw a trend, but because he made a more fundamental judgment: the underlying logic of electrification would inevitably switch from silicon to wide-bandgap materials.
This judgment was ahead of its time, but the cost of being ahead was loneliness.
In 2013, Global Power Technology launched its 4-inch production line, creating China's first market-oriented SiC device production line, achieving scaled sales in 2014. However, in those days, customers buying SiC diodes were mostly driven by scientific curiosity or small-batch trials for government projects; there was no talk of industrialization.
Those years were lived counting on fingers. The production line couldn't stop; stopping and restarting required re-conditioning the process. Orders were sporadic—one sample today, a small batch tomorrow—not enough to support scale, but every piece had to be made carefully. The team explained "what is silicon carbide" to clients while returning to continue adjusting parameters and improving yield.
No trends, no subsidies, no hot domestic substitution stories to tell—only the sound of equipment running and sweat trapped in cleanroom suits. They endured like this, stabilizing the 4-inch process, conquering automotive-grade certification, and accumulating the first batch of clients truly willing to use domestic SiC.
At the end of 2019, Global Power Technology made a critical decision: placing the 6-inch mass production line in Liuyang, Hunan. Beijing handled R&D and headquarters, while Hunan handled large-scale manufacturing. This "front shop, back factory" layout became almost the standard paradigm for Chinese hard-tech companies later, though it wasn't common in 2019. After the Liuyang 6-inch line gradually ramped up, its current annual capacity reached 100,000 wafers. From 4-inch to 6-inch, the size increased by only two inches, but behind it lay the complete reconstruction of entire line equipment, process recipes, and yield control.
Because SiC material is inherently harder to handle than silicon: substrates are exorbitantly expensive, epitaxial defect density is high, ion implantation must be performed at ultra-high temperatures, and gate oxide reliability is the Achilles' heel of MOSFETs. Every size upgrade meant tackling these problems all over again.
Betting before the trend arrives is called foresight; surviving before the trend arrives is called capability. Global Power Technology first chose the former, then proved the latter.
The rhythm of financing also reveals the company's character.
In October 2021, the Series D round closed, with SK Hynix entering through its Wuxi subsidiary. A Korean memory giant investing across tracks in a Chinese SiC device company sparked much speculation at the time: Was Global Power Technology' technical route validated? Or did overseas industrial capital need a local partner in China to hedge geopolitical risks? There was no official conclusion, but it at least indicated that in the eyes of industrial capital, Global Power Technology was not a "storyteller."
Subsequently, Oriza Holdings and TCL Ventures followed suit; in 2023, the Series E round raised hundreds of millions of CNY, involving the Beijing High-Tech Fund and Hongtai Fund; in March 2026, the Series E+ round led by Ruili Investment was completed. Over more than a decade, the shareholder list included overseas industrial capital, government-guided funds, and market-oriented VCs.
Capital never pays for sentiment; it only pays for "being earlier and more accurate than others on industrial nodes."
Returning to the matter of 8-inch wafers. Everyone in the industry knows that moving from 6-inch to 8-inch increases effective area by 78%, nearly doubles chip output, and reduces costs by 35%. These three numbers look great on a PPT, but the yield ramp-up for an 8-inch SiC production line is a deadly hurdle.
Wolfspeed itself was tormented by 8-inch yield issues in its Mohawk Valley fab for nearly five years, with capacity releases consistently falling short of expectations, causing its stock price to go on a roller coaster. Domestic players like Sanan Optoelectronics, Silan Microelectronics, and Times Electric also have their 8-inch lines in the ramp-up phase, with very few achieving stable mass production. Global Power Technology plans to reach full capacity in 2028, meaning there are still more than two years of equipment move-in, process debugging, and customer verification ahead. If any link in the chain breaks, the timeline for full capacity may slip.
More worth examining closely is Global Power Technology' product strategy.
While the entire industry fights tooth and nail in the red ocean of 650V to 1200V automotive MOSFETs, Global Power Technology bet its chips on the high-voltage platform of 2000V to 6500V.
In May 2026, Global Power Technology launched a full series of 2000V, 3300V, 4500V, and 6500V SiC MOSFETs in one go, targeting scenarios such as solid-state transformers, grid equipment, rail transit, and data center high-voltage DC, factually refuting rumors that "Global Power Technology only understands low-end diodes."
The key to this choice lies in avoiding direct confrontation: The automotive track requires facing off against Infineon, onsemi, and ROHM head-on, while also competing fiercely with domestic players like Starpower Semiconductor, Times Electric, and BYD Semiconductor. Global Power Technology' volume does not rank in the first tier on this track. But the high-voltage track is different; customers are concentrated among State Grid Corporation of China, CRRC, and the two major power generation groups. Certification cycles often take three to five years, but once entered, orders remain stable for eight to ten years, creating extremely high barriers to entry.
This is a narrow gate; the entrance is tight, but the path widens significantly once you pass through.
Looking at Shunyi again. This is a national-level specialty industrial cluster for small and medium-sized enterprises in third-generation semiconductors recognized by the Ministry of Industry and Information Technology, gathering over 30 upstream and downstream enterprises (nearly 35 in recent years), forming a pattern of "R&D in Beijing, mass production in Shunyi, application nationwide."
Previously, what this cluster lacked most was the link for large-diameter SiC wafer manufacturing. The 8-inch facility built by Global Power Technology supplements not only its own capacity but also serves as a process validation platform for the entire cluster. Domestic crystal growth furnaces, epitaxy, etching, and ion implantation equipment previously lacked nearby large-diameter trial run scenarios; now they have them.
This ecological significance is longer-term than how many chips Global Power Technology sells itself.
Around Shunyi, NEV vehicle capacities of Li Auto and BAIC are within the one-hour industrial circle of the Beijing-Tianjin-Hebei region; Xiaomi Auto is also within the scope of Beijing. State Grid headquarters is in Beijing, and AI computing centers are accelerating deployment in Hebei, Inner Mongolia, and around Beijing. Anyone who makes power devices knows what being close to customers means: fast sampling, fast response, and fast joint debugging.
By locating the 8-inch line in Shunyi, Global Power Technology compresses the "customer-production line" short link that an IDM should have into the time it takes to drink a cup of coffee.
But between vision and capacity lie real money and time. The challenges facing Global Power Technology remain severe.
First, investment in an 8-inch production line is a bottomless pit; the initial 400 million CNY is just the beginning. Equipment procurement, process R&D, and yield ramp-up require continuous funding. Although Global Power Technology has raised multiple rounds, compared to peers with valuations often reaching tens of billions, its financial reserves are not abundant.
Second, automotive-grade certification remains an insurmountable hurdle. Global Power Technology' 6-inch line recently obtained IATF 16949; the 8-inch line must go through the automotive-grade process again. From AEC-Q101 certification to OEM designation, it won't take less than two to three years.
Third, although high-voltage MOSFETs are a good track, the market capacity is still an order of magnitude smaller than automotive SiC. Whether it can support the long-term capacity of 100,000 wafers depends on when new scenarios like SST and AIDC high-voltage DC truly explode.
Looking back at Global Power Technology' fifteen years, there is a detail rarely mentioned.
In 2014, Chen Tong said in an interview that SiC mass production provided China with a latecomer advantage, and the opportunity point lay in those next few years. At that time, domestic NEVs were still surviving on subsidies, and the cost of SiC devices was five to six times that of silicon IGBTs. This statement was indeed ambitious.
Ten years later, today, SiC has accelerated from flagship option to mainstream models on 800V platforms; supercharging stations are massively switching to SiC solutions; data center power supplies are evolving from 48V architectures to 400V/800V high-voltage DC; and solid-state transformers are entering pilot and commercial phases in grids, rail transit, and computing centers. The opportunity point he spoke of back then was not "those next few years," but more than a decade later, now.
This is the cruelty of the hard tech industry. Judging the direction correctly doesn't mean you get the dividend; you also have to survive long enough until the wind truly blows.
Over the past fifteen years, hundreds of companies in China claimed to make SiC devices. Some disappeared after raising Series A, some were acquired by listed companies, some pivoted to modules and solutions. Those who truly went from 4-inch to 6-inch and then to 8-inch, stepping on every industrial node along the way, can be counted on one hand.
Global Power Technology is not the most dazzling among them, but it is indeed the most enduring.
Last year, a friend visited Global Power Technology' 6-inch wafer plant in Hunan. Upon returning, she shared her deepest impression: "Mr. Chen really maximized the cost-performance ratio of the factory utilization."
My understanding is: In a semiconductor factory, she saw the scarcest thing in manufacturing, which is pretense-free authenticity. No pretending to be an international flagship, no pretending futuristic narratives, just making wafers one by one, using every inch of the factory fully. This is the clumsiest yet most skilled form of excellence in semiconductor manufacturing: squeezing process value out of every square meter of cleanroom.
This kind of excellence isn't eye-catching, but it is particularly real. It explains why Global Power Technology has survived for fifteen years and strides forward confidently: The early4-inch plant in Beijing drilled out process stubbornness; the 6-inch plant in Liuyang forged technology and efficiency into a system; only then did the 8-inch plant in Shunyi dare to speak of large-diameter leaps.
The Chinese story of silicon carbide is not told by the most expensive buildings, but by wearing out cleanroom suits, grinding down process parameters, and pushing production lines inch by inch from nothing to something.
The topping out of a factory building is worth writing an article about for any Chinese hard tech company. But Global Power Technology' uniqueness lies in representing a type of company rarely spotlighted: not chasing trends, not telling stories, not riding hot topics, but polishing away year by year on one material, one device, and one production line. Such companies are not the hottest in the capital market, but on a nation's semiconductor industry map, they are the true foundation.
The 8-inch factory has topped out, but this is not the moment to open celebratory wine, because the ticket for the next exam follows immediately. In the coming two years, when equipment moves in, when processes run smoothly, when yield climbs above 95%, and when the first automotive-grade designation is secured—each question is harder than the concrete building itself today.
The industry doesn't need everyone to become Infineon. A healthy ecosystem needs both giants defining standards and invisible manufacturers thickening and stabilizing devices under the shadows of giants.
*In July 2023, Hunan's first all-SiC PV-storage-charging intelligent supercharging station went into operation in Liuyang.
The topping out of this factory in Shunyi is not the end, nor even the start. It is a delayed test paper: testing whether 8-inch yield can cross the life-and-death line, whether high-voltage MOSFETs can enter the BOM of grids and computing power supplies, and whether the cash flow from Liuyang's 6-inch line can sustain the loss period of Beijing's 8-inch line.
Chen Tong said chip R&D is a marathon, but I feel manufacturing is even more so. You run past one 42-kilometer mark after another, only to find the finish line is still further away.
Fifteen years ago, in that factory in Haidian, a few people faced a row of equipment, holding a breath in their hearts.
This breath was held from 4-inch to 6-inch, and from 6-inch to 8-inch.
This breath was defiance against the notion that China couldn't do silicon carbide.
Each inch crossed meant another round of equipment readjustment, process regrinding, and yield climbing from zero. Holding this breath until today, it finally exhaled from the chest, settling into the SiC wafers and devices flowing through the production lines in Liuyang and Shunyi.
China's fifteen years of silicon carbide haven't lacked storytellers; they've lacked people who turn stories into wafers. They haven't lacked PPTs planning hundred-billion-yuan production lines; they've lacked the clumsy effort of filling standard factories inch by inch, wearing out cleanroom suits, and grinding yield up bit by bit.
Fifteen years are enough for an industry to walk from obscurity to the forefront, and enough for a company to walk from solitary courage to composure.
References:
[1] Global Power Technology
[2] Hunan Daily