As featured in Fierce Electronics
Last week, the US Senate passed the USICA (The United States Innovation and Competition Act) which includes the CHIPS for America Act. Despite what the title of the USICA and its sub articles might suggest, the policy is largely a manifesto for dealing with a rising and highly competitive China with particular concern for the Western rival’s ambition to achieve semiconductor self-sufficiency and Huawei’s ascension as the leading 5G technology vendor.
The Act presents several concrete policy measures to diminish China’s access to US semiconductor technologies and 52 billion USD in federal funding to build a more resilient semiconductor supply chain. But will these policies help the US achieve what the title of the USICA suggest; improve US semiconductor supply chain resiliency and competitiveness in 5G? Will it stop the technological advancement of China and its digital economy long enough for the West to tame it?
If we were to venture to say yes to the former question, the devil is in the details. The answer to the later question is a highly complicated no.
The semiconductor industry and 5G are immensely complex topics that policymakers without industry and technical background will undoubtedly struggle to grasp in a 30-minute security or economic briefing. There are several misconceptions circulating in today’s trade and national security policy discourse that put into question the level of consideration of important nuances of the semiconductor industry and 5G.
First, the notion that the United States is not already competitive and innovative in the semiconductor industry and 5G is misguided. Intel, AMD, Nvidia, Qualcomm, Apple and others continue to set the bar for advanced semiconductor design across all conceivable device categories and enjoy outsized market share across strategic segments such as data center, 5G and emerging AI computing. Furthermore, US semiconductor technologies and devices have a tremendous global footprint lending to US technological leadership and influence in the industry.
Second, there is a very good reason why leading-edge manufacturing of semiconductors is predominantly located in Asia. It boils down to proximity to the largest manufacturing base of the global electronics industry, China, where most of the devices that use leading edge silicon such as smartphones, PCs, and servers are assembled and made. Despite Washington’s recent efforts to redirect supply chains away from China, China’s role as a global epicenter of electronics manufacturing has largely remained intact and will not likely change materially in the near future.
It is important to acknowledge that this “tech” supply chain did not form overnight. It is decades in the making. China’s role in the supply chain rapidly expanded when it opened its economic development zones to Asian Tigers economies South Korea and Taiwan providing a seemingly endless supply of low-cost labor for their burgeoning electronics and tech industries.
Those same Asian Tiger economies were previously the destination of Western semiconductor and electronics companies looking to offshore their manufacturing to take advantage of low-cost labor. This multi-decade series of offshoring has deposited deep layers of economic interdependencies that underlie today’s global semiconductor and electronics supply chain.
Third, much like their Asian Tiger predecessors, Chinese ICT vendors such as Huawei and ZTE are subject to the now antiquated belief that they win simply by undercutting their Western peers on price. In recent years Chinese ICT vendors have advanced well past their “copycat” phase and are now setting the power-performance bar for the industry prompting formidable Western competitors such as Ericsson to play catch up. This was evidenced earlier this year when Huawei out did Ericsson by a kilogram shortly after the Swedish ICT giant announced its 20 kg massive MIMO radio offering. Huawei had beaten Ericsson by 15 kilos the year before.
Fourth, leading edge process node does not necessarily mean competitive advantage in end markets such as data center and 5G in a post-Moore’s Law epoch. As we continue to move toward 2 nm, newer process nodes become economically impractical for device categories undergoing gradual if not accelerating commoditization. Intel is a great example of an American chip company that has managed to remain the global market leader by innovating on their XPU platform and advanced packaging prowess despite being almost two generations behind TSMC and Samsung in their manufacturing process technology.
Finally, the topics of semiconductor supply chain resiliency and competitiveness come off as a broader concern about tech-competitive Asia, not just China. China’s semiconductor industry is arguably a decade away from presenting any form of leadership challenge to the United States. The reality is that Taiwan’s TSMC and South Korea’s Samsung pose the most significant and immediate competitive concern regarding semiconductor manufacturing leadership. Ironically, the US government will need the expertise of either or both TSMC and Samsung to build out the onshore advanced node capacity spelled out in the CHIPS for America Act.
The Bumps and Pitfalls Ahead
There is no argument that it is important for the US government to support its industries and invest in global technological leadership. However, the previous and current administrations’ policies and hardline approach toward China present a broad range of risks that even close US allies don’t seem willing to assume in full. This cautious sentiment was on display at the recent G7 Summit in Cornwall UK. German Chancellor Angela Merkle and French President Emmanuel Macron called for a balanced approach to China who they consider a rival but also an important partner.
One of the most prominent and ongoing risks facing US industries is a prolonged chip shortage. The chip shortage has been attributed to pandemic demand shocks and insufficient chip manufacturing capacity. However, there is also the matter of massive strategic stockpiling of chips by many Chinese electronics firms that landed on the US BIS (Bureau of Industry and Security) Entity List. Ironically, US semiconductor firms have been provided licenses to export chips to blacklisted Chinese firms who have been building up their safety stock. This defensive behavior has contributed to the severe tightening of the global supply of chips especially those manufactured at legacy process nodes.
Unfortunately, the chip shortage afflicting US industries is a problem that investments in domestic fabs will not solve in the near term. Additional capacity from Intel’s fab build outs in Arizona and the introduction of their new Intel Foundry Services will not come online for another three years. It is also not likely that these new fabs will provide production capacity for legacy chips which have been the primary culprit of the current chip shortage.
Second, the US chip industry is facing growing risk of being excluded from China over time. The process is already under way as Chinese firms are designing out US technology and components to de-risk their products from US export controls. This remediation will take time but could prove problematic for the US semiconductor industry with leading firms relying on China for a quarter of their global revenue and an outsized proportion of their annual revenue growth in recent years. In fiscal 2021, China contributed 3.9 billion USD to Nvidia’s coffers versus 3.2 billion USD from the US market. In 2020, China became AMD’s largest geography generating 2.33 billion USD versus 2.29 billion USD from the United States.
The exclusion of US firms from the China market and technology supply chain will undoubtedly have impact on US technological leadership. In the long-term the influence of US technologies will diminish if the semiconductor industry becomes divided into a US camp and a China camp. This would be a significant net loss for the US semiconductor industry which has enjoyed decades of cost benefits, global market share and footprint, and a large and fast-growing market for their products in China.
Third, recent US policy toward China’s tech industry has prompted the Chinese government to accelerate its efforts to achieve semiconductor self-sufficiency. Self-sufficiency and eventual technical parity with the US are not a matter of if but when. For the moment, the Dutch government has put ASML’s license to export its leading edge EUV (Extreme Ultraviolet) lithography equipment to China’s largest foundry, SMIC, on hold at the behest of the US government. However, SMIC has made progress with DUV (Deep Ultraviolet) equipment at 7nm. The company continues to accelerate its efforts and investments to scale out production at the 28nm and 14nm Goldilocks nodes to support domestic and international demand for 5G infrastructure equipment, consumer and industrial IoT applications, and data center compute.
Fourth, the blacklisting of Chinese firms has, on occasion, inadvertently stunted US engagement in global technology standards setting collaborations. In June of last year, the US Commerce Department had to amend restrictions it imposed on US firms from engaging with Huawei. The placing of Huawei on the Entity List had the unintended consequence of US firms’ avoiding participation in 3GPP (3rd Generation Partnership Project) workgroups that define technical specifications for 5G for fear of violating technology exchange restrictions with Huawei and other sanctioned Chinese firms.
Fifth, US export restrictions imposed on friendly economies have threatened to strain traditional US trade and security alliances. Though Taiwan has joined the US State Department’s Clean Network program, US export restrictions barring TSMC from manufacturing chips for Huawei’s HiSilicon prompted the Taiwanese semiconductor industry, which imports 90 percent of its semiconductor manufacturing equipment, to initiate a proposal to incubate its own manufacturing equipment industry and localize production.
This move would not only diminish US influence on the Asia region’s semiconductor supply chain, but it would also put the future business of US semiconductor manufacturing equipment companies at risk of losing their biggest customers over time. For perspective, the loss of China and Taiwan would eliminate more than 50 percent of Applied Material’s revenue in 2020.
The final risk that policymakers should contemplate is the high likelihood that billions of dollars in government subsidies to rip and replace Chinese equipment from existing networks will not move the needle much on perceived security risks. This is because many operators who have traditionally worked with Chinese ICT vendors have legacy 2G, 3G, and 4G networks ladened with Chinese equipment. Even if Chinese vendors are restricted from supplying 5G buildouts, the new generation networks will have to coexist and communicate with legacy networks for the foreseeable future.
What’s more, the ripping and replacing of Chinese equipment adds a tremendous cost burden on operators around the globe who are struggling to monetize their investments in 5G spectrum and network roll outs. For many, the cost is too high. In September of 2020, rural service providers in the US informed the FCC that it would cost 1.83 billion USD to rip and replace Huawei and ZTE gear from their networks. In December of last year, US lawmakers approved 1.9 billion USD to reimburse affect operators.
The Red Pill or the Blue Pill?
How do you fortify US competitiveness, address national security concerns while nurturing mutual economic and trade benefits with China? Is the best path forward the prevailing continuum of acrimonious escalations and tit-for-tat retaliations that only seem to worsen US-Sino relations and threaten to undo global collaborations in critical technologies such as 5G that have evangelized US technology leadership globally?
Fortunately, there are opportunities and avenues for US policymakers and lawmakers to advance US technology and economic leadership while working out the broad ranging geopolitical differences and trade issues between the US and China. It will take a significant change in tone and mindset to divert course to a less divisive and contentious path. It is obvious that the cost of not considering a more open and amicable approach has already been significant for everyone.