That’s So SiC: China Aims to Master an EV Chip You Haven’t Heard Of

About the Series

China’s recent tech pivot marks a dramatic departure from the stellar run of the private sector-driven platform economy over the past long decade. In this new era, hardware reigns supreme. Toward that end, Beijing is already realigning investments, talent, and institutions, manifest in the very visible hand of industrial policy.

While much attention has been devoted to China chasing the technologies of tomorrow—from advanced chips to quantum computing—what’s equally important to Beijing is securing the technologies of today.

The latter is the focus of our “Tech Pivot in Practice” series because tech supply chains aren’t built with frontier technologies but rather with relatively mature, non-leading-edge technologies that have immediate commercial applications.

These technologies are crucial inputs into end products that few pay attention to because they don’t make the front pages. Yet these are the arenas in which China can dethrone market incumbents without having to invent the future. It simply has to deploy the present at scale to outcompete rivals on cost and market share.

The pivot’s success is far from assured and ultimately rests on the technologies, companies, and people behind it. This series examines the pivot’s peril and promise through these prisms.

The first of the series looks at material science’s rising importance in the tech pivot as China aims to master the silicon carbide power chip. In the second case, we examine the miniaturization of hardware, particularly cameras, and China’s bid to become the dominant supplier to the automotive industry.

Introduction

These days, an inordinate amount of attention focuses on chips at the frontier, whether it’s on ever decreasing nanometers or on how China’s SMIC might close the gap with Taiwan’s TSMC and South Korea’s Samsung.

What flies under the radar, however, are the workhorse chips that aren’t at the leading edge but are poised to see wider applications and rising market share.

Silicon carbide (SiC) power chips are a prime example. These chips are quite literally under the hood, as they are essential for controlling the flow of power in an electric vehicle (EV). Not only are these chips attractive to the automotive industry, their superior properties relative to the prevailing silicon power chips also make them optimized for use in the infrastructure needed to support EVs (e.g. charging stations, smart grids, and energy storage).

One leading indicator of SiC’s potential came in 2017, when Tesla switched to it in the Model 3. The move set off what is now expected to be a broader shift from silicon to SiC power chips, potentially making them a key component in the decarbonization drive in transport and energy infrastructure.

That market potential has not gone unnoticed in China. The development of the SiC chip made it into China’s 14th Five-Year Plan, which means it will be a recipient of Chinese state support. Like other targets of Chinese industrial policy, the SiC chip is a product for which China has much domestic demand but little domestic supply. Indeed, as the largest EV market in the world, China currently relies predominantly on foreign suppliers for SiC chips.

While these dynamics make the SiC chip ripe for throwing capital and people behind it to achieve some notion of “self-reliance,” China’s semiconductor industry has also been a cautionary tale on the limits of state largesse.

Take the recent scandal in China’s $51 billion “big fund” for chips. The misuse of central government funds led to heads rolling all the way up to the Ministry of Industry and Information Technology—ironically the main practitioner of industrial policy. That wasn’t the only instance of funny business in China’s semiconductor industry either. A fledgling chip startup HSMC apparently swindled billions from the Wuhan government in recent years.

What has become clearer to industrial policy mandarins in Beijing is that what’s core to core innovation like chips isn’t just capital, but human capital and commercialization. Throwing good money after bad products or those that lack market potential has not been a recipe for success.

But SiC chips will likely be different. For one, there is obvious product-market fit for SiC chips, with the EV industry likely to switch en masse. Two, SiC isn’t a new technology, as the technical bottleneck is largely about scaling—and China has a good track record of scaling the living daylights out of existing technologies. Three, there’s more runway for new Chinese entrants to compete because current market leaders are still struggling to meet demand.

This case study is the first in a “Tech Pivot in Practice” series that examines Beijing’s bet on SiC chips and whether public-private partnerships (PPPs) can help China overcome the talent and commercialization challenges that could hobble the pivot.

Future case studies in this series will analyze technology areas where the intersection of firms, market, and state can shed light on the promise and peril of China’s march toward mastering core innovation.

The BaSiCs: A Ready-made Product-Market Fit

While internal combustion engines require gas lines to supply it liquid fuel, EV motors rely on inverters to control the flow of electricity and cooling throughout a vehicle. Power chips are integrated into the inverters to manage its functions (see Figure 1).

Figure 1. Inverters Control The Flow Of Electricity in EVs

Note: Inverter is part of the power electronics controller.
Source: US Department of Energy.

Most power chips today are standard silicon, but using SiC semiconductors in the inverter can improve power management efficiency, prolong battery life, and reduce the EV’s weight by relying on a smaller battery. For instance, SiC chips enable the standard 75-kilowatt-hour (Kwh) battery in a Tesla Model 3 to be replaced with a 68 Kwh battery without compromising on range.

These benefits are a result of innovation in material science that has produced superior properties in SiC, chief among them durability and conductivity at higher heat. SiC has more strength per unit of area so they’re like the chip version of reinforced concrete. Moreover, SiC chips can tolerate heat up to 400°C, about 2.5x higher than standard silicon.

These properties make SiC more optimized for applications in EVs, with heat management a perennial concern because batteries get very hot. Moreover, because SiC chips work better at higher voltages, that also incentivizes EV manufacturers to switch from 400v to 800v inverters that offer better performance in terms of efficiency, faster charging, and greater range. For example, by using a 900v inverter with SiC chips in its Air Dream luxury sedan, EV startup Lucid claims it can achieve a range of over 500 miles and charge up to 300 miles in just 20 minutes, the fastest of any car. 

Although using all SiC chips in an inverter (dozens are needed) would cost roughly $500 more than using standard silicon, that could be offset by cost savings elsewhere. For instance, integrating SiC chips into higher voltage inverters could lead to the removal of more than 100 pounds of copper wiring. That could save the EV manufacturer hundreds if not thousands of dollars and reduce the EV’s weight.

A chip with superior performance that saves the EV manufacturer on copper? That’s an easy sell, and it’s little wonder that major players like Tesla, BYD, Hyundai, and Toyota are all opting for SiC chips in their latest models.

This industry-wide switching is expected to support the SiC market’s growth. According to estimates, SiC’s market value in China, the world’s largest EV market, will be near parity with standard silicon chips by 2030 (see Figure 2).

Figure 2. SiC Expected To Challenge Traditional Silicon Chips in the China Market

Note: IGBT = insulated-gate bipolar transistor, which is basically the technical term for traditional silicon.
Source: Goldman Sachs “The Green Technology Cycle SiC Equity” Report.

While the benefits of switching to SiC chips are obvious, scaling is the main constraint. In fact, supply-side bottlenecks suggest that these estimates of SiC chips’ growth are more conservative than they would be without such constraints.

These constraints on reaching economies of scale, a result of SiC’s comparatively immature and complex production process, will likely lead to demand and supply mismatches as soon as 2023 (see Figure 3).

However, this projection may need to be taken with a grain of salt because it doesn’t account for new Chinese entrants that are aiming to overcome the scaling bottleneck and narrow this gap.

Figure 3. Global Demand for SiC Wafers Set to Far Surpass Supply

Source: Canaccord Genuity.

The Scaling Bottleneck

The main obstacle to scaling SiC chips is making ever larger SiC wafers, which make up 47% of SiC chips’ cost (see Figure 4). Wafers are the vinyl records shaped discs on which chips are built—in other words, the “petri dish” of the chip fabrication process. They’re made from slicing SiC crystal ingots, which are grown under very high heat.

Figure 4. SiC Wafers Are The Most Costly Stage In SiC Chip Production

Note: substrate = wafer.
Source: Karlsson, O. & Robertsson S., (2022). Securing supply of silicon carbide semiconductors. Report No. E2021:145.

While the process of manufacturing pure silicon wafers has been mastered, SiC wafers are more difficult for a number of reasons. For one, slicing SiC ingots into wafers is tougher because they are both thinner and harder—the 50% silicon and 50% carbon composition make SiC one of the hardest materials on earth, second only to diamond. Slicing SiC ingots takes between 10-20 times longer than regular silicon ingots.

Two, growing SiC crystal ingots requires much higher heat than regular silicon (1,200 centigrade vs. 2,000-2,500 centigrade for SiC wafers), which means specialized furnaces are needed. Such furnaces are generally developed in-house by the SiC wafer manufacturer rather than being readily available on the market.

Three, handling these astronomical temperatures requires different techniques for SiC crystal growth that are less reliable compared to silicon ingots. This also tends to lead to lower yields for SiC due to a higher defect rate.

In addition to constraining volume, these factors make it more difficult to increase the size of wafers, with the current frontier at eight inches (generally, a two-inch increase in the wafer size translates into twice as many chips on a single wafer). In comparison, today’s most common silicon wafer is ~12 inches in diameter.

The limitation on wafer size inhibits the ability to scale up to meet growing demand. For instance, even global SiC leaders like US-based Wolfspeed and II-VI are just on the cusp of mass producing 8-inch wafers, with the more mature 6-inch wafers likely to dominate the market for the foreseeable future.

While increasing wafer size is a time-consuming and iterative process that requires considerable capital, it is also not rocket science. It is not nearly as difficult as reducing nodes at microscopic levels on the most advanced chips that a company like TSMC produces.

For example, Chinese firm Sanan Optoelectronics already announced plans to begin production of its first 8-inch SiC wafers at its Hunan plant by 2024. While commencing production and scaling production will still be separated by a number of years, the catch-up cycle for SiCs is nonetheless faster relative to the most advanced chips. This is likely a key reason behind China’s focused approach to carving out significant market share in SiCs.

China Aims To Climb the Ranks of SiC Manufacturing

China is still an insignificant player in SiC chips for the time being, but it has several things going for it (see Figure 5). One, it will have extant domestic demand for these chips because it is the world’s largest EV market. Two, it is adept at scaling existing technology and driving down costs. Three, it’s an area where industrial policy might prove useful. Four, the political mandate of reducing reliance on foreign technology supply chains adds impetus to developing a semiconductor product where China appears to be closer to the current frontier.

Figure 5. Chinese SiC Wafer Suppliers Want to Crack Foreign Dominance

Source: Yole; company estimates for 2021.

Investments are already pouring in, as China’s demand for SiC wafers is forecasted to more than double from around 7% of global demand in 2019 to 16% by 2025. The capital is mainly coming from Chinese companies riding the domestic EV market growth, with the likes of NIO, Xpeng, Li Auto, and BYD among the investors (see Figure 6). Even telecom giant Huawei, eyeing potential for SiC in its 5G infrastructure business, has gotten into the action.

Figure 6. SiC Investments Largely Correlate with China’s EV Production

Note: Data includes plug-in hybrid EV and battery EV.
Source: China Association of Automobile Manufacturers; ITJuzi; Qianzhan Industrial Research Institute.

In short, there doesn’t appear to be a shortage of private capital in SiC production. The potential addressable market, as well as the lack of strong market incumbents, make SiC chips a more promising investment than many of the other core innovation verticals that China has identified.

But China still has some catching up to do to get to 8-inch wafers. And here’s where a nudge from industrial policy in the form of PPPs might be effective in closing the technology gap.

State Innovates, Private Sector Commercializes 

Unlike in other chip verticals where Chinese government largesse would throw money at flagship projects, such as the now embattled chips “big fund,” the state is instead leveraging the talent in its universities and labs to support the private sector’s drive to scale up SiC production.

Although this is a reversal of China’s usual industrial policy practices, it is an approach similar to how WolfSpeed, current SiC market leader, was spun out of a university research institute in North Carolina. Founded by a group of researchers at North Carolina State University in 1987, WolfSpeed was formed to commercialize their discoveries in SiC technology.

It appears Beijing hopes that a souped up and deliberately focused PPP model will spawn a dozen WolfSpeeds. That’s because state research institutes and universities hold considerable patents on SiC manufacturing that can boost yields and lower defect rates, as well as patents on integrating SiCs across various applications from EVs and renewable energy to infrastructure.

Since 2012, Chinese players have become some of the most prolific filers of SiC patents, which have been traditionally dominated by Japanese companies (see Figure 7). Various institutes under the Chinese Academy of Sciences, as well as Shandong University and Xidian University, have played instrumental roles in SiC research and development (R&D) and have even jointly filed patents with companies. This is unsurprising considering the greater concentration of Chinese STEM doctorates in academia.

Figure 7. China Dominates SiC Patent Filings after a Decade

Source: KnowMade Power SiC June 2022.

So, instead of technology transfers from foreign entities, China’s own state labs and research institutes can execute technology transfers to the private sector to commercialize. This allows private companies to devote resources to commercialization and scaling—an emphasis that’s currently lacking among core innovation firms—because they can essentially outsource R&D to state institutions.

One such example is SICC, a company spun out of China’s Shandong University, a leading material science institution owing to the late Jiang Minhua, who is largely credited as the “godfather” of SiC in China.

After Jiang passed away in 2011, his work’s commercial potential was still not realized. While many of his colleagues were reluctant to pick up that slack, one of Jiang’s younger contemporaries from nearby Qilu University, Zong Yanmin, was less reticent. Zong seized the opportunity to purchase Jiang’s patents with his own capital and created SICC with himself controlling an 80% stake.

By 2021, SICC already was making 40,000 6-inch SiC wafers annually and aims to quadruple the volume to 300,000 in five years across its operations. After listing on the Shanghai Star Market in 2021, the company is now valued at over 45 billion yuan ($6.3 billion).

SICC isn’t an isolated case either. Of the 30 key projects targeted for scaling SiC production to ~4 million wafers per year by 2026, over half of these projects have been supported by state and academic institutions via talent or technology transfers (see Figure 8).

Figure 8. State-backed Research Drives Chinese SiC Production Expansion

Note: Historic data on state-supported production unavailable. Capacity estimates based on companies’ announced plans and may not reflect actual production.
Source: Semiconductor Material Industry Association; China Business Information Network; Oriental Fortune Securities Research Institute; company announcements.

Ramping up the PPP model is ultimately aimed at replicating a tech manufacturing ecosystem that China has achieved in sectors such as batteries and solar panels. That ecosystem, in Beijing’s vision, would center along the Shanghai-Suzhou corridor to create something of a “silicon carbide valley.” For instance, the University of Shanghai for Science and Technology just attracted two Australian-Chinese returnees in material science to bolster talent in this area.

While it took WolfSpeed around a decade to transition from 6-inch to 8-inch wafers, Chinese speed hopes to cut that time significantly, as already demonstrated by a company like SICC (see Figure 9). Not only does China’s PPP model leverage top talents from state institutions to shorten that development cycle, having the world’s largest EV market means that there’s strong incentive for the private sector to scale this technology to meet demand.

Figure 9. SICC’s Catch-Up Cycle Was Shorter Than Current US Leaders

Source: Author calculations.

If all goes accordingly, China will likely accelerate SiC development and become a leading player in the power chip market. Such success would notch the Chinese leadership a major “win” under the banner of technology self-reliance because China will have mastered a chip that is increasingly crucial to the energy transition in transport.

AJ Cortese is a senior research associate at MacroPolo. You can find his work on industrial technology, semiconductors, the digital economy, and other topics here.