Cognitive Dissidents

Demystifying Chips, US-China Relations, & Intellectual Property

Mon Nov 20 2023

supply chain diversificationgovernment policiesAI regulationsemiconductor industryEVsbattery supply chainsUS-China relationsexport controlsgreen technology

Description

This episode explores the challenges and complexities of diversifying supply chains away from China, the impact of government policies on AI regulation, and the future of the semiconductor industry. It also delves into the strained relationship between the US and China, the role of export controls, and the battleground of EVs and green technology.

Insights

Supply Chain Diversification

Companies face challenges in diversifying their supply chains away from China due to limitations, workforce availability, and geopolitical influences. Moving manufacturing to other countries can be challenging, and export controls add complexity to decision-making processes.

Government Policies and AI Regulation

Governments are questioning optimized supply chains and increasing regulatory efforts regarding AI. The trajectory of US-China relations in AI regulation is uncertain, and finding common ground between different regulatory systems is challenging.

Semiconductor Industry Challenges

China's advantage in the green tech and critical mineral space gives them leverage over the US. Other countries like India face challenges in building their semiconductor capacity, while export controls and government interference impact the industry.

Impact on EVs and Battery Supply Chains

Export controls and industrial policies affect the semiconductor industry, creating challenges in supply chains and market access. The EV space is becoming a battleground, with Chinese companies aiming to dominate EV battery supply chains.

Chapters

Introduction
Supply Chain Diversification
Challenges of Diversification
Government Policies and AI Regulation
Regulatory Challenges and US-China Relations
US Policy and Export Controls
Semiconductor Industry Challenges
Challenges in Semiconductor Manufacturing
Impact on EVs and Battery Supply Chains
The Future of EVs and Green Technology

Summary

Transcript

Introduction

00:00 - 07:03

Jacob Shapiro introduces Paul Triolo, an associate partner for China and technology policy lead at Albright Stonebridge Group, as a guest on the podcast.
They plan to discuss semiconductors and artificial intelligence.
Jacob expresses his admiration for Paul's insights on technology and technology policy.
They briefly mention Evie's neck as a topic of discussion.
Jacob thanks Paul for accepting the invitation to be on the podcast.
They talk about the use of social media and its potential decline in popularity.
Jacob mentions that he has been a fan of Paul's work for a long time and expects there to be plenty to talk about during the episode.
Jacob asks Paul about his thoughts on de-globalization, multipolarity, near-shoring, and the breakdown of US-China relationship in relation to tech supply chains.
Paul discusses concerns about supply chain dependencies and concentration of manufacturing capacity before and during the pandemic.
He mentions that governments are becoming more aware of supply chain importance and implementing industrial policies as a result.
The US-China relationship has led to discussions of decoupling or reducing dependence in high-tech sectors due to national security interests.
Jake Sullivan's outline of technology policy between the US and China implies some level of decoupling or reduction of dependence in computation, biotech, and green tech sectors.
Concerns about supply chains dovetail with broader trends among governments regarding resilience, decoupling, and de-risking.
The US and EU prefer the term "de-risking" over "decoupling" due to their highly entangled economies.

Supply Chain Diversification

06:41 - 13:44

Companies are constantly evaluating and managing risks in their supply chains.
The level of government intervention in industries, supply chains, and markets is a topic of debate.
Different industries have different approaches to diversifying their supply chains away from China.
Moving advanced manufacturing from China to other countries can be challenging due to limitations and workforce availability.
Before the US-China tech competition and the pandemic, companies were already looking at alternatives to China for manufacturing.
Apple's supply chain concentration in China makes it difficult for them to fully move production to other countries like India.
The geopolitical landscape is influencing companies' decisions on diversifying away from China.
Export controls and industrial policies add complexity to companies' decision-making processes.
European companies manufacturing in China face challenges in dividing up their supply chains without creating excessive costs.
Japanese companies seem more open to maintaining dual supply chains with one specific to China.
Infrastructure plays a critical role in deciding alternative manufacturing locations, with China having a well-developed infrastructure compared to other countries like India.

Challenges of Diversification

13:15 - 19:55

China has optimized its manufacturing sector for the past two decades, making it difficult for companies to diversify away from China.
Customers are asking companies to diversify away from China, but it is a challenging task.
Major disruptions to supply chains, such as export controls, can impact companies' business plans and require them to find ways around these challenges.
Moving manufacturing to Mexico may protect intellectual property, but there are concerns about violence and random acts of violence in Mexico.
Intellectual property protection depends on the industry and location. Companies like Nvidia are good at protecting their core IP.
Companies focus more on the risks associated with concentrated business areas and how to handle crises related to geopolitical tensions.
Many companies are now requesting crisis management exercises to prepare for potential scenarios involving geopolitical issues.
The South China Sea and Taiwan Strait are areas of concern due to their importance in global shipping. Disruptions in these areas could have significant impacts on businesses.
Companies are trying to mitigate risks and find ways around disruptions by considering alternative options like moving production to Mexico or diversifying supply chains.

Government Policies and AI Regulation

19:27 - 25:50

Global supply chains were optimized in a free environment with the free flow of capital, goods, and people.
Governments are now questioning these optimized supply chains due to vulnerabilities.
The concentration of advanced semiconductors is mainly in Taiwan, which was a result of market optimization.
The US is facing challenges in implementing industrial policy due to lack of experience.
There is a lack of understanding and experience in dealing with emerging technologies like AI and crypto within governments.
Governments are rushing into addressing supply chain dependencies without fully understanding the complexities involved.
Jake Sullivan published an article emphasizing the need to work with China on containing risks related to artificial intelligence, but there are doubts about the potential for collaboration given past experiences.
Both the US and China have increased their regulatory efforts regarding AI in the past year.

Regulatory Challenges and US-China Relations

25:26 - 31:55

China has rushed out generative AI regulations and is willing to engage both bilaterally and multilaterally with the US on this issue.
There are divided opinions on whether the West should decide first on regulating AI before including China, or if excluding China would hinder the development of an interoperable regulatory system.
China has been invited to the UK AI Summit for industry discussions but not to the government part of the summit, which may affect their decision to attend.
The US is putting strict controls on Chinese companies' access to advanced GPUs, which could impact their ability to develop large-language models.
There is a challenge in finding common ground between different regulatory systems, censorship, data concerns, etc., in order to engage in meaningful discussions about AI governance.
China wants a full seat at the table and feels they should be included since they have a significant capacity for generative AI development.
The trajectory of US-China relations regarding AI regulation is uncertain, with conflicting signals from leaks and policy actions.
Recent controls imposed by the US seem softer than expected, leading to confusion about whether the US is softening or tightening its stance.
South Korea and Taiwan have received exemptions while facing policy challenges from the US. The uncertainty raises questions about future carve-outs or changes in rules under different administrations.

US Policy and Export Controls

31:26 - 38:32

US policy in the technology space, particularly regarding China, was outlined by Jake Sullivan last year.
The three parts of the Sullivan Doctrine include maintaining a lead over China in key technologies, controlling sensitive technology with restrictions, and considering advanced compute, biotech, and green tech as strategic force multipliers for national security.
The goal of US controls on China's advanced compute and AI capabilities is to limit their development and freeze them at a certain level.
However, China already has a significant advanced compute capacity with many supercomputers on the top 500 list.
Export controls are being used to restrict China's access to advanced semiconductors like GPUs and their ability to manufacture them domestically.
The assumption that US companies have unique technology that can be locked away from China is false as technologies constantly evolve and there are multiple ways to achieve similar results.
Chinese companies have found ways around restrictions, such as using existing tools to manufacture semiconductors like Huawei Mach 60.
The goal of stopping China's AI development through export controls is unclear and not supported by all allies.
Export controls have unintended effects, high collateral damage for US companies, and are difficult to enforce with leakage and workarounds.
From the Chinese perspective, these controls are seen as an attempt to contain their technology development rather than just focusing on military purposes.
Incentivizing Chinese companies to innovate by restricting their access to foreign chips may lead to erosion of market share for US tool makers over time.
This situation could escalate into a tit-for-tat retaliation between the US and China in terms of control over critical minerals and supply chains.

Semiconductor Industry Challenges

38:13 - 44:46

China has a significant advantage in the green tech and critical mineral space, which gives them leverage over the US and US companies.
The US and China are both learning from each other's efforts to control technology, creating a complex situation for companies caught in between.
The strained relationship between the US and China could lead to increased conflict, potentially affecting technological advancements globally.
Apart from the US and China, other countries like India are also trying to build their own semiconductor capacity but face challenges due to weak infrastructure and lack of ecosystem support.
Building a semiconductor fab is a complicated process that requires significant investment, customers, and an established market.
India's attempt to enter the semiconductor industry faces difficulties due to its lack of manufacturing history and expertise in advanced technologies.
To succeed in the semiconductor industry, India needs to focus on building an ecosystem with R&D capabilities, education systems, and infrastructure before considering establishing fabs or factories.

Challenges in Semiconductor Manufacturing

44:19 - 50:56

Building a fab or factory for semiconductor manufacturing requires all the necessary pieces to make commercial sense and recoup the investment.
India is spending public money on semiconductor projects, but it's a gamble as they need to ensure the projects will be successful in the long run.
The US is inviting established companies or subsidizing them to build semiconductor facilities, reducing the risk compared to India.
Japan has decided to focus on high-end technologies like two nanometers and advanced packaging instead of competing with mature node companies.
Japan is investing billions of yen in building a factory called Rapidus, which is also a gamble as they don't know who their customers will be or if the technology will succeed. They aim to develop an education system and hardware manufacturing culture.
Morris Chiang, founder of TSMC, is skeptical about the long-term commitment of the US and India towards advanced manufacturing.
The economics of semiconductor manufacturing are tough, requiring a clear technology base and commercial roadmap. India faces challenges in building this culture.
The future growth promised by artificial intelligence, Internet of Things, and 5G networks relies on having enough chips. However, export controls and other factors make it difficult to ensure chip supply for these industries.
The majority of the semiconductor industry focuses on mature or legacy nodes used in cars and equipment controllers. Advanced chips for phones and AI training are critical but narrow parts of the industry.
Companies need a solid income base from legacy chips to invest in R&D for advanced chips. Only a few companies design advanced semiconductors while others manufacture them (Foundry Business).
The chip shortage was caused by poor supply chain management by auto companies during the pandemic. Capacity was shifted from cars to more advanced products due to higher margins.
There isn't a shortage of capacity for mature nodes globally but it may be distributed unevenly. China also plays a significant role in producing mature node technology.
Export controls complicate the semiconductor industry, as some US companies rely on Chinese manufacturers like SMIC for mature node technology.

Impact on EVs and Battery Supply Chains

50:37 - 57:33

The export controls on SMIC, a semiconductor company, are complicating its relationship with US companies that rely on its mature node technology.
There is a shortage of advanced GPUs used in large language models, but the market is expected to solve this problem as Taiwan builds more packaging capacity and other players enter the market.
Generative AI and large language models are still in the early stages, but enterprise-level applications are starting to emerge in China.
The demand for hardware to train large language models remains high, particularly for GPUs, despite potential business losses for NVIDIA in China.
The semiconductor industry is affected by export controls and industrial policies imposed by governments, which can disrupt supply chains and market access.
It is uncertain whether government interference will negatively impact the semiconductor industry or if it will lead to overcapacity or restricted markets.
Building fabs in the US has proven challenging due to cultural reasons and a lack of trained workers at competitive costs.
The EV space is becoming another battleground with Chinese companies aiming to become global players in EV battery supply chains.

The Future of EVs and Green Technology

57:07 - 1:00:18

EVs (Electric Vehicles) and battery supply chains are the next big battleground in the automotive industry.
Chinese companies are becoming major players in the global EV space and have significant control over the supply chain.
EVs use significantly more semiconductors than traditional cars, as they are becoming more like computers.
The topic of EVs is intertwined with US-China relations and has a global impact, particularly in Europe.
There is a need to be cautious about politicizing green technology, as it has broader implications for climate change and the future of the planet.