Chip Policy for the Green Transition
Diving into energy ministries and public research agencies’ efforts supporting chips for climate tech
Happy Earth Day week!
Chip Capitols is celebrating our planet by breaking down the different ways governments incorporate semiconductor research and manufacturing into their environmental sustainability and green transition plans.
We have already begun a series of deep dives into the unique characteristics of major economies’ big-picture chip incentives (see China’s here and Europe’s here). This article will take a magnifying glass to these policies to examine how the US, EU, and China incentivize semiconductor R&D relevant to energy grids, smart cities, the industrial automation, and more.
In the US, green chip policy spans the technology readiness level (TRL) spectrum from basic research to prototyping and manufacturing, with the Department of Energy’s national laboratory ecosystem playing a central role. European governments’ efforts are limited to the earlier stages of green chip research, and this article will use France to demonstrate how member states incentivize green chip research within strict EU frameworks. Lastly, China incentivizes green chip research under a unique model discussed previously on Chip Capitols called jiēbǎng guàshuài (揭榜挂帅), roughly translating to “revealing the lists and taking command.” More explicitly than the US or EU, the PRC also seeks to incentivize chip manufacturing in line with its green transition needs, but, as discussed previously on Chip Capitols, China is yet to coordinate its local incentive policies under a single national agenda.
[This article is a bit longer than usual, so I would suggest reading in a browser.] Today’s agenda covers:
Semiconductors’ footprint and handprint,
Environmental and supply chain goals inspiring green chip policies, &
How American, French, and Chinese government agencies use chips to implement national green transition goals.
The Footprint and Handprint of Semiconductors
Chipmakers describe the environmental impact of semiconductors as divided between a footprint and handprint. On the negative side, the footprint includes greenhouse gases and chemicals emitted in the wafer fabrication process, as well as the energy required to run the world’s compute capacity. This footprint is constantly shrinking, and, in the grand scheme of the green transition, the environmental cost of chips pales in comparison to the green digitization chips enable. Chips’ handprints include their role in enabling smart energy grids, the Internet of Things (IoT), and greater industrial efficiency. The second subsection below will dive deeply into the segments of the chip industry, namely Wide Bandgap (WBG) technology in discrete and analog semiconductors, most critical to the the green transition.
Footprint
Computing for the information and communications technology (ICT) industry accounts for slightly over 2 percent of global carbon emissions. Even before chips begin sucking power in massive data centers or on laptops, however, semiconductor fabs rely heavily on electricity, water, and chemicals to run a plethora of enormous machines etching the finest designs on minuscule sheets of silicon.
An excellent 2020 study by researchers from Harvard, Facebook, and Arizona State University looked at the various activities contributing to the ICT industry’s overall carbon footprint. As we will see shortly, chips’ increased operational efficiency due to more advanced structures and algorithms has greatly lowered one part of the ICT carbon footprint, but carbon emissions from semiconductor manufacturing have been more stubborn. The share of life-cycle carbon emissions due to hardware manufacturing increased from 49% for the iPhone 3GS to 86% for the iPhone 11. Though this measure includes manufacturing for parts other than semiconductors, chipmakers’ carbon emissions are considerable. With the world having emitted 35.26 billion tons of carbon in 2020, TSMC contributed 15 million tons (0.043% of global emissions), Intel contributed 2.88 million tons (0.008% of global emissions), and Samsung at 29 million tons (0.092% of global emissions; this includes emissions from Samsung’s non-chip businesses as well).
Though there remains significant room for improvement on the manufacturing side, the operation of semiconductor-powered computing activities has witnessed a remarkable increase in energy efficiency.
Data centers are enormous storage sites for connected computer servers used to store, process, or distribute large amounts of data. Since 2010, the number of Internet users has doubled, and internet traffic has grown 12-fold. Nonetheless, aggregate energy use at data centers has remained largely flat over the same time-period because the energy intensity of data centers has decreased 20 percent annually. That data centers’ aggregate energy use has remained flat —rather than decreasing on net— should not itself be alarming either because these central depositories manage a growing share of consumers’ overall compute as they transfer data from personal devices to the cloud. As a result, data has migrated from relatively low efficiency servers (personal cellphones and computers) to highly efficient centralized servers.
Individual consumer devices such as mobile phones and PCs have become significantly more energy efficient due to new chip technology and better software. In 2017, the US’s over 3.4 billion consumer electronics accounted for 4 percent of the US’ total energy consumption, down by 14 percent from 2013.
Transmission networks stand between data centers and consumer devices. Per a study of studies, the energy intensity of transmission networks declined by a factor of about 170.32 from the year 2000 to 2015. As data becomes increasingly mobile with the rise of IoT, the energy efficiency of data transmission is critical to mitigating carbon emissions.
Though semiconductor manufacturing still has significant work left to mitigate its energy consumption, compute efficiency has seen significant improvement alongside the march of Moore’s Law. The Decadal Plan for Semiconductors, authored by the Semiconductor Research Corporation in cooperation with leading chipmakers, has identified compute efficiency improvement as one of the “seismic shifts” facing the industry, suggesting further footprint reductions to come.
Handprint
As important as lowering semiconductors’ environmental footprint may be, the handprint of chips has a transformative impact many times greater than any of the footprint effects described above. Semiconductor-enabled digital technologies have the potential to lower greenhouse gas emissions by 15 percent, nearly a third of the reductions required by 2030 to keep global temperature rise under 2°C.
Whereas the excitement around artificial intelligence and consumer gadgets revolves around advanced logic and memory semiconductors, green transformation of the industrial, energy, and automotive sectors relies on the mainstream Discrete, Analog, Other (DAO) category of semiconductors. Comprising two-thirds of these heavy industries’ chip inputs, DAOs are semiconductors that transmit, receive, and process real-world information, such as temperature, light, gravity, and voltage. Discretes include diodes and transistors that switch electrical power on-and-off, and analog chips that enable power management tools and devices dealing with radio signals. The “other” category covers optical sensors used in cameras to sense light, as well as a host of other chips found in IoT devices.
Critically, many sectors of the green economy operate under extreme heat and pressure conditions that could degrade the silicon in ordinary chips. To that end, Wide Bandgap (WBG) semiconductors allow devices to function efficiently at the higher processing frequencies required in 5G base stations, heavy home appliances, electric vehicles, energy grids, and high-speed trains. For example, WBGs operating at high voltages reduce the amount of electrical power wasted in automobiles, thereby reducing vehicle fuel consumption and carbon emissions.
Innovations in silicon carbide and gallium nitride as alternatives to pure silicon for chips make WBG applications in emerging clean-tech industries possible, but WBGs face challenges impeding widespread adoption. Being a relatively nascent technology, inputs for silicon carbide and gallium nitride are not as readily available as those for conventional chips. Semiconductor equipment manufacturers also need to develop new measuring and packaging tools as this equipment has to-date only been developed to work with silicon.
Continued development of DAO semiconductors, specifically WBGs, has the potential to transform countless existing industries. The industry envisions 5G as opening the gates to further transformations in smart heating, smart transport, smart manufacturing, and smart agriculture to reduce greenhouse gas emissions equalling that of 81 million cars. Companies also predict that optimizing water usage with smart sensors could reduce US water consumption by 450 billons within the decade, equaling the water use of 4 million American households. As we will see, governments around the world are eager to realize these transformations, and semiconductors loom large in the sustainability plans they have put out.
Three Approaches to Green Chip Policy
American, European, and Chinese green chip policies function against the backdrop of a global supply chain for DAO semiconductors. These chips are made using mature technology, with the wafer manufacturing for conventional and WBGs used in the energy sector accounting for 7 to 76 percent of the global capacity for mature nodes (those ranging from 28 to 180 nanometers). Europe, Japan, and the US held 68% of global DAO manufacturing capacity in 2019. However, China’s lower cost of manufacturing is helping the PRC raise its share of the DAO sector to a forecast 49% by 2030 —from 17% in 2019.
National green chip policies address technological and manufacturing needs within the global supply chain context. This section summarizes the role of semiconductors in energy and sustainability programs in the US, EU, and China.
The US Goes Industrial
Semiconductors have played a central role in US energy policy since long before the CHIPS and Science Act of 2022 made chip subsidies front-page news.
Multiple administrations of both parties have affirmed that microelectronics are critical to technologies underpinning energy systems, digitization, and productivity enhancements to existing industries. The Department of Energy has also long conducted research and development projects that “extend across the RDD&D [TRL] continuum—from the development of novel materials and technologies to prototyping and commercialization of these technologies.” The agency conducts this research through its network of 17 national labs, which house key facilities, talent, and intellectual property.
The US has a long history of using DOE to advance research in sensors, wired/wireless communications, power control, and computing for energy efficiency applications; a more recent phenomenon is the Biden administration DOE’s emphasis on supply chain security. Following a February 2021 executive order to multiple agencies, directing them to analyze US supply chains, DOE identified semiconductors as posing a key vulnerability to the “Energy Sector Industrial Base.” DOE’s subsequent plans to address these weaknesses expanded the agency’s traditional focus on semiconductor R&D to a more supply chain–focused strategy of “strengthening the U.S. position in energy-efficient semiconductor design, fabrication, and advanced packaging, as well as in advanced power electronics.”
Underlining the emphasis on semiconductors as a tool for energy supply chain security, the CHIPS and Science Act included $500 million in authorizations to implement the bipartisan Micro Act. Overshadowed by the over $52 billion in semiconductor manufacturing and research appropriations, this provision directs DOE to establish dedicated Microelectronics Science Research Centers for the R&D and commercialization of chips critical to the energy sector.
The EU Chugs Along with Longstanding Green Chip Research
As described in a recent Chip Capitols article, European Union rules on state subsidies are extremely strict, though the likely-to-be-soon-passed EU Chips Act is monumental in its loosening of these restrictions for critical fab construction projects. As a result of historically tight state-aid rules, EU green chip policies have focused almost exclusively on low-TRL (technology readiness level) basic research.
EU member states’ public research programs for semiconductor applications in the energy sector function primarily as Important Projects of Common European Interest (IPCEIs). Set out in 2014, these rules set out the criteria under which member states may financially support R&D for innovative projects, and the EU Commission has identified microelectronics as one of six Key Enabling Technologies that generally qualify for these projects.
A 2018 IPCEI proposed by France, Germany, Italy, and the UK (then an EU member) demonstrates how European governments conduct research into green chip technology. With the four countries providing a combined €1.75 billion in public funding, the project focused on five technology areas, all of which hold key nexuses to either reducing chips’ footprint or growing chips’ handprint:
Energy Efficient Chips: The countries proposed to develop new solutions to improve chips’ efficiency in order to lower the energy consumption of end-use devices, such as cars.
Power Semiconductors: Plans to develop new chip components for smart appliances and electric vehicles served to increase the reliability of final devices.
Smart Sensors: The plan identified new optical, motion, and magnetic field sensors as key to improving electric vehicle safety.
Advanced Optical Equipment: The countries sought to build upon the EU’s already strong position in semiconductor manufacturing equipment by developing more advanced tools for future high-end chips. (This pillar is not as specific to energy-sector applications as the others.)
Compound Materials: The plan focused on developing new compound materials to be used in lieu of silicon and on developing devices suitable for manufacturing non-silicon chips. This directly addresses the challenge mentioned earlier as limiting WBG technology from becoming more widespread.
Beyond the IPCEI framework, the EU recently invested massively in technology for the green transition through the Horizon Europe program. Unlike the US’ recent efforts, however, Horizon Europe did not expand the EU’s energy policy to focus on semiconductor supply chain issues. It contained a relatively small €35 million line-item funding research into chip design and manufacturing for low-energy consumption, connectivity, sensing, and power management. Critically, eligible projects must address “low-TRL research.” Despite its small size, Horizon Europe is significant in that it indicates the EU will not soon abandon its tight restrictions on subsidies for manufacturing —no matter how dire energy-related chip supply chain concerns become.
China’s (Theoretically) Two-Pronged Plans
Like the US and EU, China has long stressed the importance of chips to its national goal of “digitization” (数字化) through greenification (绿色化). This agenda theoretically extends to government support for both chip research and manufacturing.
However, as we have discussed multiple times in previous Chip Capitols articles, China’s central government struggles to unite local governments behind a national chip strategy. As a result, centrally administered semiconductor research programs regularly prioritize funds for chip projects relating to green tech, but locally administered manufacturing funds do not prioritize projects based on national needs.
Speaking on a central government–level, China has for several years prioritized reducing semiconductors’ carbon footprint. The Guiding Opinions on Strengthening the Construction of Green Data Centers (2019) focused on reducing data centers’ energy usage, and the subsequent Three-Year Action Plan for the Development of New Data Centers (2021) expanded this footprint reduction policy by focusing on the energy use of telecommunications infrastructure connected to China’s data centers.
After years of focus on shrinking the footprint of compute, the 14th Five-Year Plan’s National Informatization Plan (“十四五”国家信息化规划) took a major step in calling for the expansion of chips’ handprint. It called for the “acceleration of research and development of… core chips” in order to promote “5G + Industrial Internet” (5G+工业互联网). As the next section will elaborate, the term “industrial internet” includes applications across IoT that are enabled by 5G. The Plan’s logic essentially flows as follows: chips —> 5G —> IoT/digitization —> greenification.
Touch It & Feel It – Green Chip Policy on the Ground
Having discussed the high-level policy goals and frameworks governing green chip policy in the US, EU, and China, this section will discuss how the three regions implement these policies. In the US, our emphasis will be on the Department of Energy; in the EU, we will use France as a case study for how member states support green chip research within EU frameworks; and in China, we will examine how lower-level administrative plans implement high-level Party and State decrees.
US DOE Infrastructure, Know-How, and Funds Supporting Research and Manufacturing
The US Department of Energy Office of Science’s user facilities provide cutting-edge tools that private sector participants in DOE programs can leverage for research. The 17 facilities have core capabilities that generally fit into one of seven categories: 1) materials & processing, 2) devices, circuits, & system integration, 3) fabrication & prototyping, 4) characterization & testing, 5) computing & algorithms, 6) manufacturing, and 7) commercial deployment
Together, these categories show the full TRL spectrum, from basic research to commercialization, of US green chip policy. Details on the categories most relevant to green technologies are listed below:
Materials & Processing: DOE’s five Nanoscale Science Research Centers offer tools for advanced materials synthesis and chemical processing. Being on the less mature end of the TRL spectrum, resultant innovations reveal how physical phenomena can enable more efficient computation, communication, and sensing.
Devices, Circuits, & System Integration: Also generally at less mature TRLs, some DOE researchers seek to establish foundational knowledge on how to tailor hardware structures to specific computational activities for greater efficiency.
Computing & Algorithms: DOE works alongside chipmakers and fabless chip designers to design new process and memory architectures that can increase the energy efficiency of artificial intelligence and machine learning algorithms.
Fabrication & Prototyping: Some DOE facilities work with industry partners to develop new semiconductor manufacturing equipment and chip technologies that companies can later commercialize.
Manufacturing: Moving up the TRL spectrum beyond basic and applied R&D, DOE also conducts research on manufacturing processes, which it uses to provide technical assistance for industrial energy efficiency and cost savings in fabs. These efforts serve to reduce the footprint of semiconductor manufacturing.
Commercial Deployment: Beyond research, development, and manufacturing support, DOE also directly supports the commercial deployment of chip technologies in critical energy applications. The DOE Loan Programs Office (LPO) provides partially subsidized direct loans to support automakers that incorporate semiconductor components. LPO can also give loans to support the early deployment in post-pilot commercial plants of new semiconductor technologies that help mitigate greenhouse gas emissions.
Multi-Pronged French Implementation of the EU’s Green Chip Framework
For the sake of brevity, we will look at how only one member state implements the EU’s IPCEI and Horizon Europe frameworks –France. The key point that will become readily apparent is that French policies strictly adhere to EU requirements that state funds only support green chip research, not subsidize manufacturing.
Having contributed nearly €900 million to the previously mentioned joint IPCEI with Germany, Italy, and the UK, France launched the Plan Nano 2022 to support semiconductor research that could feed into automotive, IoT, and other increasingly digitized downstream applications. In addition to technological sovereignty and social inclusivity, one of Plan Nano’s key pillars is the “launching of new innovation projects serving the digital and energy transitions.” This pillar includes “power electronics for improving the efficiency of electric vehicles and energy-efficient embedded AI technologies” [translations mine].
Unlike France’s permanent national research agency, the Agence nationale de recherche (ANR), Plan Nano works closely with industry to fund joint research. This is legally possible due to the state-aid exemptions provided by the EU’s IPCEI framework, which serve to leverage private investment in support of EU-guided and member state–implemented technology goals.
Beyond IPCEI, France’s CEA-Leti (Commissariat à l'énergie atomique: Laboratoire d'électronique des technologies de l'information) operates within the framework of Horizon Europe to research silicon carbide and gallium nitride technology for WBG semiconductors. Since 2014, Leti has offered its facilities and personnel in partnership with industry to shrink transistors in WBG chips. These R&D results serve to increase energy efficiency in electric vehicles and other industries relying on motor controls and large power supplies.
Chinese Ministries Implement Digitization and Greenification Goals, One Action Plan at a Time
Last year featured the first notable implementation of China’s aforementioned 14th Five-Year Plan’s National Informatization Plan. “In order to deeply implement the major decisions and deployments of the [Chinese Communist] Party Central Committee and the State Council,” six administrative departments issued the Industrial Energy Efficiency Improvement Action Plan (2022). The document emphasizes leveraging semiconductors’ handprint to increase sustainability across other industries. The Action Plan proposes three steps for “greenification” (绿色化) through “digitization” (数字化):
Digital Twin Systems (孪生系统): The Action Plan calls for R&D in cloud computing and artificial intelligence to “digitally track, centrally manage, and intelligently schedule energy consumption and carbon assets.” It calls for industry, research institutions, and universities to conduct joint research into how such methods can apply to existing heavy industries.
Industrial Chain Efficiency: The Action Plan calls for “targeted enablement schemes” in collecting and analyzing energy use across various industries’ supply chains, both upstream and downstream. It suggests using centralized data management and certification to identify and solve inefficiencies.
Resource Library of Solutions (解决方案资源库): The Action Plan lastly suggests building a “resource library of solutions” to help industries implement typical case studies of energy efficiency improvement through digital solutions. It hopes that such a resource would foster development of an “Industrial Internet + Energy Efficiency Management" ecosystem (“工业互联网+能效管理”生态).
What technologies does the central government want public funds to prioritize? And what industries most need digitization through greenification? In 2021, the Ministry of Industry and Information Technology (MIIT) published two catalogues to answer just that.
The Recommended Catalogue of Energy Saving Technology Products for the National Communication Industry (国家通信业节能技术产品推荐目录) lists technologies with critical applications in industrial digitization, offering varying degrees of detail to define what they actually are. Listed technologies include “green and low-carbon data center series energy-saving technology,” “AI energy management systems,” “smart DC low-carbon green data center solution,” “intelligent temperature control systems,” “efficient modular integrated cold station in data center,” “spray liquid cooling edge computing workstation,” among many others.
The National Catalogue of Industrial Energy Conservation Technology Recommendations (国家工业节能技术推荐目录) complements the Energy Saving Technology Products Catalogue by discussing how identified technologies can be applied to well-established industries to lower their carbon emissions. Sectors it calls out for digitization include steel, non-ferrous metals, building materials, petrochemical and chemicals, energy storage and renewable energy utilities, and industries producing residual heat and pressure.
Lastly, the National High Technology Development Program (国家高技术研究发展计划; the 863 Program) has played a critical role awarding funds for developing semiconductor technologies in a range of energy efficiency (先进能源技术领域) applications. One 863 Program funding opportunity offered ¥80 million (11.6 million USD) for projects researching high voltage, silicon carbide semiconductors to be deployed in advanced energy fields. Another 863 Program funding opportunity offered ¥300 million (43.6 million USD) for projects researching new materials technology for efficient semiconductor lighting.
Being centrally administered, China’s research award programs can channel significant funds toward energy-focused semiconductor research projects. It remains to be seen, however, whether the latest high-level plans will succeed in bringing the PRC’s provincial and municipal integrated circuit investment funds in line with national priorities. Doing so will be critical to expanding China’s green chip policy beyond research to include DAO manufacturing with a focus on green technology clients.
Shrinking Footprints and Growing Handprints
Preventing the worst of climate change requires the fundamental transformation of industrial sectors, and semiconductors underlie the technologies that can make these transformations happen.
Though they disagree on many other fronts, the US, EU, and China are in lockstep on the need to exploit the handprint of digitization, while minimizing the carbon footprint of computation. The regions take different approaches to what I have coined “green chip policy,” and these differing techniques yield different results.
More than in any other area of chip policy, government policymakers should learn from their counterparts’ successes and mistakes in developing and deploying semiconductors to decarbonize polluting industries. Supply chain security will continue to be a sensitive political game, but any region’s success in the digital and green transitions will get the planet one step closer to averting climate catastrophe.