On December 18th, 2020, SMIC (Semiconductor Manufacturing International Company), China’s largest foundry, was added to the Bureau of Industry and Security’s (BIS) Entity List. This was not a surprising move by the US Department of Commerce given the increasing stringency and aggressiveness of the Trump Administration’s tech trade policy toward China over the last couple of years, especially with the advent of the Coronavirus Pandemic. Besides, Secretary Wilbur Ross indicated in September that SMIC would likely join Huawei on the BIS’s Entity List. What are the implications on the semiconductor industry and the future of 5G?
To date we have seen US bans instituted against a rapidly growing list of Chinese tech firms. These measures have been met with US tech firms scrambling for licenses to continue their business with their Chinese customers. In most instances, US firms have been allowed to sell to their Chinese customers who collectively represent the growth market for the US tech industry, in particular, the semiconductor industry. In September, Intel was provided a license to continue to sell its products to Huawei. Shortly after, Qualcomm was granted a license by the Department of Commerce to sell 4G modem chips to Huawei.
The recent “blacklisting” of SMIC is a significant advancement of the US Department of Commerce’s campaign against Huawei and other Chinese tech companies motivated by national security concerns and consternation over China’s growing leadership in 5G globally. More specifically as it pertains to SMIC, the idea of China achieving semiconductor self-reliance is perceived as an existential threat to US technology leadership and tech industry competitiveness. This likely explains why BIS clamped down particularly hard on SMIC by denying by default the export of any US technologies to the Chinese foundry that could be used in sub-10nm process node manufacturing.
According to the BIS press release, the blacklisting came about amid concerns of SMIC alleged ties with the Chinese military and their supplying advanced silicon in support of China’s ambitious military build up and modernization. The measure against SMIC was also likely intended to block any remaining channels for Huawei’s HiSilicon to access advanced node manufacturing for mobile chips used in Huawei’s premium P Series smartphones, which have been the flagship of the company’s Consumer Business Group (CBG).
If the goal of the US Department of Commerce is to hobble Huawei’s supply chain or disarm their 5G business, the Department of Commerce has to recognize that this endeavor will be a tricky matter with many nuances. Effective China tech containment policies will require more than restricting HiSilicon’s access to advanced fabs whether they are SMIC or TSMC. It will require a more holistic understanding and treatment of how semiconductor technologies and the industry have and continue to evolve.
One important question to contemplate is what is semiconductor leadership? Outside the semiconductor industry, leadership is perceived as CPU-centric. Whoever has the processor with the fastest clock speeds, largest cache size, highest number of cores wins. The semiconductor industry tends to see itself in terms of traditional market segments of memory, CPU, GPU, analog, ASIC etc. These views are dated and need refreshing as the nature of computing is quickly evolving.
The reality is that all categories of silicon are coming together in some combination with new integration techniques, technologies and architectures to create a system. This system/platform trend in the semiconductor industry has gone mainstream with Intel evolving beyond their CPU-centric legacy to becoming what CEO Bob Swan calls an XPU company. Despite their 10nm manufacturing slip, Intel is proving that you don’t need devices produced on the most advanced node to deliver competitive or superior systems or platforms.
This system/platform perspective is not unique to Intel. Arm has been toeing the system and platform lines across their ecosystem for quite some time with the idea of heterogenous compute. The likes of Apple and Qualcomm are taking chip designs for mobile and PC computing to the next level especially with the introduction of the M1 processor for the Mac as well as Qualcomm’s Fused Ai architecture for its 800 series premium SoCs. It is little wonder that Nvidia and AMD are working on their own XPU plays with their recent acquisitions of Arm and Xilinx respectively. At the end of the day, the system matters.
Why is this system/platform trend important? In this nascent age of ubiquitous AI, applications are becoming hybrid applications that run on some form of fused AI and general purpose compute. We have seen many of these applications emerge on the smartphone. For example, a fast-growing range of computational photography applications blend ML operations with ISP functions to automate complex computational techniques on device as well as create entirely new photographic functions such as computational HDR (High Dynamic Range) for photos and videos.
It is becoming increasingly obvious that Huawei will tap into new hybrid compute models and innovate in distributed AI architectures for edge computing to sustain the competitiveness of their 5G carrier products and solution offerings. In fact, one can argue that Huawei has considered alternative architectures long ago that take advantage of their vertical integration up and down the 5G technology stack; an advantage that recent excitement in the US and UK regarding open RAN technologies and standards by the O-RAN Alliance and OpenRAN will be challenged to dispel over the near and mid terms.
“Bottom line, we need to think about computing differently as well as revisit the benchmarks and standards that determine “critical” or “strategic” technologies.“
While the US government’s bi-partisan effort to persuade allies to ban Huawei and ZTE technologies from their 5G deployments was initially met with broad pushback from allied governments and operators alike, the “The Clean Network” program spearheaded by US Secretary of State Mike Pompeo has gained traction in recent months especially in light of the pandemic. The initiative has managed to curtail Huawei’s 5G contracts outside of China from 96 wins at the beginning of 2020 down to 12 contracts as European countries expressed varying levels of Clean Network commitments.
The banning of Chinese firms from 5G networks may be warranted from US national security and foreign policy perspectives, but they are not likely to prevent China from pressing forward with their ambition to become the leader in the 5G era or to acquire semiconductor self-sufficiency in the near future. The current course in US tech trade policies will most certainly drive China to innovate and invent ultimately at the expense of the US semiconductor industry which will lose their market share in the massive and fast-growing China market.
The US-centric technology trade policy of the Trump Administration has not been without backlash in the global technology community. It has seeded some possibly unintended consequences that could subvert the US government’s current and future efforts to control China’s access to advanced technologies. For one, the Taiwanese government recently moved to incubate a local semiconductor equipment industry in a move to afford itself independence from US trade policies that might restrict their opportunities with the Mainland. November of last year, the RISC-V Foundation moved its headquarters to Switzerland over fears of increasing US trade curbs.
Despite the Trump Administration’s hardline efforts, China has become the crucible for 5G innovation and the hotbed of growth in terms of connections and deployments with over 600,000 5G sites live across over 300 cities. China’s big three operators are bringing 5G-enabled solutions to market through their collaborations with enterprise customers across industries including transportation and logistics, manufacturing and utilities. Many US technology firms that neXt Curve engaged with in 2020 attested that China quickly recovered their tech investment during the course of the global pandemic and managed to drive substantial innovation and progress in the industrial application of 5G as well as AI computing over the last 12 months.
In the past couple of years we have seen China redouble its efforts to become semiconductor self-reliant as trade and geopolitical tensions with the US continue to rise. On the investment front, SMIC went public on the Shanghai Stock Exchange raising over 50 billion yuan for the Chinese foundry in July of this year. On the technology front, SMIC announced last month that its new FinFET N+1 process node is equivalent to TSMCs 7nm process. Moreover, Huawei has outlined alternative mobile computing concepts and architectures that use edge computing offload and distributed AI approaches that could help them work around some of the constraints they will face with access to leading edge semiconductor technology and capacity.
As the semiconductor industry continues to grapple with the economics of the limits of Moore’s Law, integration, and packaging technologies and new architectures for designing systems on chip are going to have equal if not more weight than scaling. Given that the vast majority of packaging and integration is done in China, Beijing might have an ace or two up its sleeve to play in this latest round in the US-China tech trade contest. This is an important understanding that US policymakers current and future should acquire if they have not already.
While the hardline against Chinese firms may play well to domestic US politics, the incoming Biden Administration should consider how the current US tech trade policy comes off to Beijing. Does it embolden the Chinese or influence desired behavior and outcomes? Even today, US firms continue to sell to Chinese companies including Huawei and ZTE. TikTok is still TikTok as we knew it as is WeChat. To the Chinese, the bans and the bluster may seem like a bunch of sound and fury that will not inspire them to fulfill their commitments to what now appears to be a defunct Phase 1 Trade Deal with China much less anything else.
The global technology supply chain is a complex, multi-layered beast that is constantly and rapidly changing. The best policies for US policymakers may not be the most obvious or the most politically convenient especially for the long term.
For more insights on neXt Curve’s semiconductor technology and end market research themes contact us for a detailed briefing with one of our analysts. Follow our site to get updates and notifications of our research and our 2020 research agenda.
Related Content & Media
- Tech Insight: Heterogenous Computing
- Qualcomm Tech Summit 2020: Key Takeaways
- Huawei MBB 2020: Key Takeaways
- Interdigital 6G Summit 2020: Key Takeaways
- COVID-19 Pandemic: the Crucible of 5G Innovation and Leadership
- Industry Insight: Apple Silicon
- Industry Insight: China’s Trade “Win-Win” and the Phase 2 Trade Deal
- Industry Insight: COVID-19 Tests China’s Resiliency
- Industry Insight: Apple’s China Problem
Managing Director, neXt Curve
December 22, 2020
This material may not be copied, reproduced, or modified in whole or in part for any purpose except with express written permission or license from an authorized representative of neXt Curve. In addition to such written permission or license to copy, reproduce, or modify this document in whole or part, an acknowledgement of the authors of the document and all applicable portions of the copyright notice must be clearly referenced.
© 2020 neXt Curve. All rights reserved.
If you would like our research articles and blog posts emailed to you, please subscribe with your email address below.