The race for semiconductor supremacy | FT Film

You can enable subtitles (captions) in the video player

SPEAKER: The slumbering giant that is America has finally awakened.

SPEAKER: Chips is now at the very heart of US-China tensions.

SPEAKER: Technological leadership has been the source of US strength since World War II.

PAT GELSINGER: We shouldn't be dependent on any one port, any one country, any one island.

RANA FOROOHAR: An economic competition that is manageable could turn into something bigger and hotter and unmanageable.

[MUSIC PLAYING]

JOE BIDEN: American manufacturing the backbone of our economy began to get hollowed out. Companies moved jobs overseas. Today, we're down to producing only around 10% of the world's chips despite leading the world in research and design of new chip technologies. But, folks, where is it written, where is it written that America can't lead the world once again in manufacturing?

RANA FOROOHAR: It's hard to think of a component, in any product really, that's more ubiquitous and more essential than the semiconductor.

CHRIS MILLER: Today, almost any device with an On/Off switch has one or often dozens or hundreds of semiconductors inside.

IGOR MIKOLIC-TORREIRA: They power our cars, airplanes, trains. Communications rely on them. Factories, ordinary businesses, and finance and banks. Money moves, thanks to these chips.

JOHN NEUFFER: The US no longer manufactures the most innovative chips. 7 nanometers and below, Taiwan manufactures 92% of those chips. 8% of those chips are manufactured in South Korea, and none are manufactured in the US. We just want more of the new manufacturing that's coming online to happen in the US.

RANA FOROOHAR: There was a sense that globalisation is inevitable. We should offshore these jobs. No big deal if the Rust Belt is hollowed out.

That decision is one of the reasons that the US elected Donald Trump, and that is why Joe Biden is pushing industrial strategy. And that is why we are at a pendulum shift in the political economy, not just in the US, but in many developed countries.

PAT GELSINGER: When people think about industrial policy, it has been a bit of a dirty word in the US politic and, hey, free markets should reign. This is not like any other industry. This is so critical for the future, and I think that's why it really deserves somewhat unique attention from an industrial policy perspective.

SPEAKER: Members of Congress I have the high privilege and distinct honour to present to you the President of the United States

JOE BIDEN: We're going to make sure the supply chain for America begins in America. That's why, that's why we came together to pass the bipartisan CHIPS and Science Act.

RANA FOROOHAR: This is really the crowning jewel in the Biden policy crown.

SPEAKER: So what's behind the CHIPS Act is an attempt to protect the economic security of the US far into the future.

SPEAKER: The whole CHIPS Act is $52.7 billion. $39 billion is for grants. There's also an investment tax credit on top of this, which gives a 25% tax credit off of any fab that's constructed or improved upon.

PAT GELSINGER: The focus of the CHIPS Act is on producing the chips that the US needs for its National Security requirements, Department of Defence, the intelligence agencies, and the associated companies that are going to build products for them.

SPEAKER: As the US looks to become a real leader in global chip making again, it's looking at creating centres across the country where it can build up on the skills, both of the people and of the local suppliers that are needed around these large chip fabs.

These are fabrication plants. These are the big factories, and they produce the wafers, large discs of silicon, on which chips are printed and then cut. So this is really at the heart of all chip making, and these plants cost 10s of billions of dollars.

PAT GELSINGER: You have some of the largest industrial projects that have ever been undertaken, and they're building the smallest things that have ever been built on Earth. And if we get there, we have good transistors and a good cost structure, we're going to win foundry customers, and manufacturing will shift back to the US. Today, the Rust Belt ends. Today, the silicon heartland begins.

[MUSIC PLAYING]

SPEAKER: Phoenix has been named as one of five US workforce hubs focused on manufacturing jobs.

JAMES LEWIS: It is really an exciting day to be here, and this builds on 40 years that Intel has been in Arizona. We put our chips on the table, right, with this investment. Now, we're going to put our shovels in the ground and get this thing underway, and we are excited for the opportunity to expand and now have over $50 billion of invested capital in the State of Arizona.

CHRIS MILLER: Arizona has been a centre of semiconductor manufacturing in the US for several decades. But in the past couple of years, it's seen major new investments, both from Intel and from TSMC, both of which are building brand new, cutting edge chip making facilities.

KATE GALLEGO: Advanced manufacturing has been an incredible industry for Arizona, and part of the reason I'm so focused on it as mayor is I feel that it will develop a healthier economy, more diverse, and that so much of the future of the world's technologies will need chips, which are going to be made here in Phoenix.

The CHIPS Act is going to change Phoenix forever. We have seen unemployment rates change. And we're also working on the workforce side, and we were recently selected by the White House as one of five workforce hubs. I'm very focused on trying to create a knowledge-based economy with high-wage jobs. That's why I'm betting on the semiconductor industry. Arizona State University has been a key partner in our semiconductor expansion.

KYLE SQUIRES: The rate at which the technology is moving, that is the smallness of the features, right, you hear about 3 nanometer, you hear about 2 nanometer. This is coming, OK? And those are really pushing the boundaries of what physics will allow. Well, how do we solve that problem? That gets an engineering student excited. Engineering and these students, they want to solve problems.

[MUSIC PLAYING]

MARK LI: My name is Mark Li. I'm from Kazakhstan. I was really interested into working in semiconductor industry. I chose material science because material science is something that people in industry are looking for. I'm interested to be an inventor and making technology in a specific area. Inventing new materials, improve the chips, trying to make an impact on our society.

Now, I feel like I'm getting very close to that dream. This is our clean room and we have a lot of tools, but you could notice one of the things you would start noticing right away. This area is white, that area is yellow. And again, coming back to when you do any sort of lithography work, when you're exposing, like, a film and you're developing it, you want to have no UV light.

IGOR MIKOLIC-TORREIRA: The US doesn't have all the talent to staff these fabs. You're going to need about 27,000 people with fairly high technical skills over the next 5 to 10 years because we're going to need thousands of people from outside the US with the right skills, which means we're going to have to facilitate the immigration of these people.

MARK LI: US has, like, such an environment where you can succeed if you want to. If you really strive for something you can achieve it. And that's what attracts me in the US. I didn't feel the same way back in my country.

PAT GELSINGER: Talent is a big challenge we need to overcome. There's two sides of it. One, we need to work harder on STEM education. That's one side of the coin, and that's going to fix the medium and long-term. The short-term, we need to have some modifications to our immigration policy, which is very, very broken for high-skilled immigration.

CHRIS MILLER: If you look at the founding of Silicon Valley there were workers from Europe, from Korea, from Egypt, for example, in the founding of some of the earliest companies in Silicon Valley.

JAMES LEWIS: It's an area of concern, it's an area of priority, and the CHIPS Act and funding some of that skills development. But at the same time, I look at this and the response that we've gotten, for instance from university system in the US at the community college level, at the University level, the top tier schools, has been phenomenal.

RICHARD WATERS: Intel is one of the most underrated stories in the tech industry, and that's a big thing to say given how prominent it's been and everybody knows the brand. But what I don't think people appreciate is that for about half a century this company was the complete leader in an entire industry. It invented integrated circuits, which are what we think of as microprocessors, as chips. And it basically led that industry and manufacturing for a half a century.

And what that meant was that with each new generation of chips it designed, it also developed the manufacturing processes to build them. And it was developing things that were smaller and faster than anybody else could design, and it was typically two to three years, or even more, ahead of the entire global chip industry. But what's happened more recently is a different story.

In the last five to eight years, Intel lost its edge, it lost its ability to just keep staying one step ahead of the rivals, if you like. And other companies caught up.

JAMES LEWIS: The plan that I've laid out for Intel says that we'll invest about $30 billion in capital per year for five years-- $150 billion. These are extraordinary investments. And when you go look at those of TSMC or Samsung, they're similar. We've been declining for 30 years. Can we start inclining for the next 30 years?

SPEAKER: Here is a world where material is grown into crystals with controlled impurities, with intolerances, which are molecular, even atomic. The objective? Semiconductors, transistors, and diodes needed by the millions.

SPEAKER: So at the very beginning of the chip age, Gordon Moore, who was one of the founders of Intel, was looking for a way to explain to people how chips were going to be in everything, how electronics, the whole world of electronics, was about to go through a revolution that nobody could quite understand.

CHRIS MILLER: In 1965, Gordon Moore set out in a paper what became known as Moore's Law. The idea behind it was that the number of transistors per chip would double every year or two, and that was just a prediction. It's not a law of nature, but it's proven correct for over half a century.

SPEAKER: The basic goal, whatever the product, make it smaller, more reliable, more economical, extend its function.

CHRIS MILLER: If you go to the store and buy a new smartphone, just the primary semiconductor on the phone will have 10 or 20 billion tiny transistors carved into the silicon, and each one of those transistors is so small. It's roughly the size of a coronavirus. And there's no device that we've produced at such vast scale using such microscopic components.

Now, as transistors get smaller and smaller, it's getting harder than ever before to shrink them further. Major companies like Intel or TSMC, they've outlined very clear pathways to keep Moore's Law alive for years to come.

JAMES LEWIS: Until the periodic table is exhausted, Moore's Law isn't finished, right? We're just going to keep bending the edge of science to find how to keep these advancements underway.

CHRIS MILLER: For the future of AI, for the future of all sorts of technological advances, it's absolutely critical that we keep this rate of innovation happening as rapidly as possible for many years into the future.

JOHN NEUFFER: Chips have been at the centre of electronics and computing from the very beginning. And it's been very easy to ignore them. They're just there. We take them for granted, but all that's changed. We've all discovered during the pandemic that an interruption of chip supplies can affect everything, can affect all manufacturing industries. So suddenly, the chip world is being turned on its head and Intel, and the all-purpose chip world is really trying to rethink its role.

SPEAKER: Six million vehicles were taken out of production plans in 2021 due to the semiconductor shortage.

CHRIS MILLER: Over the course of 2021 and 2022, car companies globally saw several hundred billion dollars in lost sales because they couldn't get the chips they needed to finish their cars.

RICHARD WATERS: Car lines shut down just because there weren't enough chips available. Car makers cut their purchases of chips thinking they would sell fewer cars, and so all the chip makers instead sold to PC companies. And then suddenly there was a whiplash. How do we de-risk? How do we make sure that we have multiple suppliers so that we aren't reliant on a single supplier? And how do we make sure there's enough capacity in the supply chain?

CHRIS MILLER: Today, 90% of the world's most advanced processors, the types of chips in your smartphone or PC, or in data centres and telecoms infrastructures, can only be produced by TSMC, which has all of its most advanced production in Taiwan.

JOHN NEUFFER: In 1990 we manufactured 37% of semiconductors here on US shores. Now we manufacture 12% and falling. Overconcentration of manufacturing in East Asia is neither good for our national security or economic security.

RICHARD WATERS: TSMC, the global leader at the moment, basically focuses most of its manufacturing close to its home and operates absolutely massive plants with huge output.

CHRIS MILLER: When the company was founded in 1987, it had a really unique business model. It would only manufacture chips, not design any in-house and that business model let it serve a wide variety of customers. Companies like Apple or Nvidia, and as a result it's grown far larger than any other chip maker of its kind. And because of its scale, it's been able to drive down costs, but also hone its technology because for every silicon wafer it produces, it gathers data, hones its production processes. And so it's not a coincidence that TSMC is both the world's largest chip maker, it's also the world's advanced chip maker.

RANA FOROOHAR: Should we have 92% of all semiconductors in Taiwan? Maybe not such a good idea. That's an incredible vulnerability, and it really makes you stop and think, how did we get to a place where countries and companies thought that this was OK? You know, I think a lot about what would happen in this country and around the world if there were to be a blockade of Taiwan. If there were to be a hot war in the South China Seas.

The only near-term comparison we have to something like this is the war in Ukraine. That created a global market event. We had inflation. We had huge corrections. We had, both in companies and countries, an interruption of the supply chains. Not just in chips, but in everything through the South China Seas would be 1,000 times that.

JOHN CORNYN: This is a bad day for President Xi and the Chinese Communist Party. The slumbering giant that is America has finally awakened to the challenge that we face from the People's Republic of China. Their aggressive posture in the region, and the potential they would have of cutting off our access to advanced semiconductors.

SPEAKER: Another day of unpredictable Chinese drills around Taiwan this morning, simulating sealing off this self-governed island. Taiwan says dozens of Chinese warplanes once again crossed the very sensitive de facto Maritime border of the Taiwan Strait.

CHRIS MILLER: Over the past decade, Chinese military power in the Taiwan Straits has grown dramatically as China's ramped up its military spending and focused almost all of its military efforts on threatening Taiwan.

[NON-ENGLISH SPEECH]

RICHARD WATERS: Imagine what would happen today if instead of an oil embargo, like we had in the 1970's, there was a chip embargo. If the economy collapses, and if there are large interruptions in everyday life and the functioning of society, you could essentially expect production lines across the world to just stop.

RANA FOROOHAR: The global paradigm is changing, and that creates bumps and that can create conflict. And there is a real risk at this moment that an economic competition that is manageable could turn into something bigger, and hotter, and unmanageable. And that really does worry me.

JAMES LEWIS: What if China, in some way, got control of Taiwan? It doesn't have to be an invasion, but a much lower budget approach would be to make sure that the right folks won the next Taiwanese election. And what happens if the Taiwanese government says to the US, thanks a lot, you can go now? Conflict need not be military conflict, and that's a point that gets lost.

We have a chance to manage that conflict using things like export controls. Using things like industrial policy, but conflict is inevitable. And that's where the semiconductor industry and the CHIPS Act come in.

RICHARD WATERS: In the tension that's emerged between the US and China, the computing industry and the tech industry is at the heart of future global competitiveness and national security. And in that tech rivalry that's emerged, chips are right at the centre. The US has realised that it's China's lack of a real skill in making advanced chips that is the only thing that's holding it back from becoming an advanced technology power. And so the US is actively acting to stop China gaining that chip technology.

ANTHONY BLINKEN: It's not in our interest to provide technology to China that could be used against us, and at a time when it's engaged in a buildup of its nuclear weapons programme in a very opaque way.

[MUSIC PLAYING]

CHRIS MILLER: Today, China is a relatively minor player in the chip industry. China spends as much money each year importing semiconductors as it spends importing oil. And China has to buy all these semiconductors from abroad because its domestic capabilities at the cutting edge are quite limited.

JAMES LEWIS: In addition to the 25% to 35% it consumes, it has another 25% to 30% of world supply chain goes through China. So it plays a critical role in the supply chain as well.

CHRIS MILLER: In 2014, Chinese President Xi Jinping identified semiconductors as a core technology that China should produce domestically. And since that point, China has poured tens of billions of dollars a year, roughly one CHIPS Act a year, into subsidising its own domestic production.

XI JINPING: [NON-ENGLISH SPEECH]

QIANER LIU: The semiconductor industry has become like a symbol of whether China can build its own self-reliance technology industry. At the centre of the government-led investment is the National Integrated Circuit fund, known as the Big fund. So Big Funds have been raising 340 billion RMB, which is $47 billion dollars since 2014.

And so far, with little fruit from such a big investment. So Beijing have to basically rethink whether this is such an approach that's useful.

CHRIS MILLER: In aggregate, Chinese chip industry still lags behind technologically despite all these generous subsidies. And when you look across the supply chain, China is years behind the cutting edge in Taiwan, Korea, or other countries. The big shift that we're seeing is more of a bifurcation between the China focused chip industry and the chip industry selling to the rest of the world.

SPEAKER: It's not just the basic science that has to be grokked, but there's a lot of tacit knowledge, if you wish, shop floor knowledge into how to make it work that's very hard to replicate. TSMC has a tremendous amount of shop floor knowledge, that tacit knowledge of their workers, of how to make those processes actually run and produce high yields.

SPEAKER: The US is considering tightening curbs on the export of chips to China.

SPEAKER: Advanced semiconductors, a key development and national security priority in China.

LESLIE WU: [NON-ENGLISH SPEECH]

JOE BIDEN: And we seek competition, not conflict, investing in our alliances and working with our allies to protect advanced technologies so they will not be used against us.

RICHARD WATERS: The chip world is now moving into specialised chips for specialised purposes. In AI, that means GPUs, a different type of chip that was originally designed for gaming, but has now been repurposed for AI. And we've seen NVIDIA, which really leads the world in GPUs, become the world's most valuable chip company, and indeed, hit $1 trillion valuation just on the strength of these single-purpose chips that work well in AI.

SPEAKER: It used to be that emerging technologies, especially in the United States, came out of the Department of Defence. Think of it, satellite systems in general, semiconductors themselves, but also things like RADAR, SONAR, GPS that now we can't live without. But because they came out of the National Security enterprise of the government, you had a chance for them to diffuse slowly into civilian life and the US government could manage whatever risk they had of their adversaries using this same technology.

Today, there are a bunch of technologies, artificial intelligence being just one of them, where the lead of this is actually happening in the private sector. And the US government is trying to catch up. And as the government appreciates what the national security implications of these technologies are, how does it now try to put rabbits back in the box, or manage the risks that these technologies have to the national security?

GINA RAIMONDO: Competing with China requires everyone on the field operating as a whole of government, and working in a bipartisan way with Congress. It requires the use of the full extent of our economic, diplomatic, and military tools.

CHRIS MILLER: In 2022, the US imposed new restrictions on the transfer of certain chips used to train artificial intelligence systems to China. Today, almost all of the world's advanced AI chips are designed by US firms and manufactured in Taiwan, and data centres in China that were training AI systems were using these chips as well.

And the US wants to limit China's access to these high-end chips with the aim of constraining China's ability to deploy AI systems for Defence and intelligence use cases.

JAMES LEWIS: Technological leadership has been the source of US strength since World War II. And that was when the US discovered that if you have the most advanced technology, you could make your own security. You could achieve leadership in the world. And a lot of what's happened since then has been an effort to keep America's technological supremacy.

RANA FOROOHAR: The US and China have been doing a supply chain dance for a couple of decades, maybe for decades now. Sometimes it's gone well, other times it's been more dysfunctional. And I think we're really kind of nearing a zero-day where we begin to understand, OK, we can't have the status quo anymore, and what are supply chains, globally, going to look like.

[MUSIC PLAYING]