America’s Big Chipmaking Blunder

This is possibly the most important machine in the world right now from a technological, economic and geopolitical point of view. Without it, the global economy would slow down the pace of technological advancement, it would stutter. They would cost $200 million each, and of the roughly 200 that exist today, only a handful are operational. American soil, what's more, successive American administrations have had to fight to ensure that nothing is sold in China, and that's despite much of the early work on the origins of the technology being done in the US. ., this is a story about where physics meets business and has a massive impact on the global economy, so how did the United States manage to miss out on this colossally important piece of technology?

It's called extreme ultraviolet lithography or euv machine and it produces the most advanced semiconductors in the world. UV technology is the only reason we have iPhones. which are as fast as they are the reason we have this AI revolution with chatbots and PT chat, the device has made the only company that produces it, the Dutch company asml, the largest technology company in Europe. Lithography is how circuit patterns are etched onto chips. a very, very important part of the semiconductor manufacturing process the scale of the extreme ultraviolet lithography machines is mind-boggling each device is the size of a bus but is designed to etch patterns into chips that measure billionths of a meter across the way they They do it is an extraordinary feat of engineering: a high-powered laser is fired at a small target of tin droplets approximately 50,000 times per second, creating a plasma that emits extreme ultraviolet light that has a wavelength of 13 1 12 nanometers.

In fact, this light does not occur naturally on Earth. It is absorbed by most materials, including air, so the entire process takes place in a vacuum and a series of mirrors reflect and focus the UV light. Each mirror is coated with layers of lithium and silicon and polished to a smoothness less than the thickness of an atom, which is important. because any imperfection reduces the quality of the chips that are produced. Halfway through this process, the Euv light is reflected off a grating containing the pattern of the circuits to be etched onto the chip, these are then reflected and focused further to create the pattern. even smaller before you get to the silicon wafer, each wafer is etched with billions of those patterns, the problem was that it is so difficult to make that the semiconductor industry is always racing to make things smaller and its The ability to continue doing so helps the global economy continue to grow now.

Of the layers of materials used to make semiconductors are one atom thick, they do not get thinner than one atom. Semiconductors eventually faced a problem where they were really pushing the limits of physics. In the 1980s, scientists began to think that ultraviolet light might be the best way to get to that atomic level, the US government through the so-called National Laboratories has always been helpful to the industry. semiconductors, in some ways supports some of these fundamental advances; In fact, the Department of Energy ended up investing tens of millions of dollars in EUV research in three laboratories in the US.

To make the leap from research to reality, an alliance was formed with companies like Intel, AMD and Motorola, there they matched any government spending in 1999 a little-known Dutch lithography company called asml joined together to remember that in the 1990s the Japanese were a major threat to American dominance of the chip industry, the US government supported asml above the Japanese companies that were really leaders in photolithography at the time asml pulled out all the stops and calculated that the technology could be commercially ready in 2006. EV technology is really a moon shot, it was incredibly expensive and incredibly complicated.

By 2012 progress had been made but more was needed as euv might actually be possible but asml actually needed a lot of investment so they turned to their customers, tsmc, Samsung and Intel decided wow! It's really important, we better put a few billion dollars into this to make sure it actually happens and intelligence was the biggest contributor to that. It was really thanks to ASML's stubbornness that it took years and billions in funding to actually produce EUV technology as we know it. Now finally in 2017, ASML begins shipping the machines in significant quantities, but despite billions of dollars of investment from American companies like Intel and the US government, the first generation of devices was for tsmc and Samsung and it wasn't because asml didn't do it.

I wanted to sell machines to Intel. Intel thought euv was really important, but like their competitors, they saw the problems it had. It was a decision they made and it is a decision that cost them enormously. Basically, then-CEO Brian Kranich didn't believe Intel could. make euv work profitably and would be fine without it, intel was the biggest city company in the industry, it dominated the industry for several decades, everyone else was trying to catch up, this is the scale on which Intel and tsmc have been able to make chips, they have both done so consistently. it got smaller but with intel leading the size reduction until 2018 when tsmc overtook intel for the first time the reason was that their core was quite simple tsmc was making chips with the euv machines and intel didn't soon after after Samsung's Galaxy Note 10 smartphone hit the market. the first consumer device with chips made using the EV process that Apple follows, but then there's a problem: Huawei, which was the world's largest smartphone maker at one point, was buying chips from tsmc, the fact that manufacturers of Chinese devices making some of the most advanced smartphones caught attention.

The US government's attention to Huawei is a very dangerous thing. If you look at what they have done from a security point of view, from a military point of view, it is very dangerous. One of the ironies of Euv is that much of the underlying technology originated in the United States, but the direct and indirect beneficiaries of that technology have not really been concentrated in the US - there was a company in mainland China that was buying chips to a company in Taiwan using equipment made in the Netherlands. This vital cog in the wheel was manufactured by a company in another country. a country they couldn't directly control, so Washington had to work hard to persuade the Dutch not to allow asml to export euv to China.

The details of how asml did not ship to China have actually been a bit murky throughout the entire process. what we know is that they claim over and over again that we never shipped euv to china in 2012. intel was 15 times bigger than nvidia and almost twice as big as tsmc. That's when it first invested in asml, but Intel's failure to move early on euv has allowed several rivals to overtake it. Intel's growth has stagnated as tsmc and especially Nvidia have prospered. Intel's leadership team under Kranich really handed over the keys to the kingdom and that caused a ripple effect within Intel.

Kranich let some of the other leaders be replaced. and there was kind of a struggle to get that company back online and one of the conclusions that we came to new leadership: we need euv, we have a very strong partnership with asml and our plans to now remain at the forefront of using euv are well underway, that is the current CEO of Pat Ginger Intel, is putting the weight of the company behind the next technology. The new thing in the euv world is called high numerical aperture or high na euv to give you an idea of ​​how important this is.

Intel has been promoting it loudly. for anyone who will listen, they have the first High Anda machine, the stakes are high. America's leadership in semiconductors and the success of President Joe Biden's chip law, which has set aside hundreds of billions of dollars in subsidies for the semiconductor industry, will allow advanced semiconductor manufacturing to return here in the United States after 40 years ago, politicians have finally realized how important semiconductors really are and Intel's entire plan to turn the company around is based on the idea that they can actually get government funding to build semiconductor plants and become leader in semiconductor technology