[ad_1]
Although microchips have become essential in many products, a small number of producers with little capacity – and hunger – to produce raw material chips, which are fundamental to current technologies, have prevailed in their development and manufacture. And because it takes hundreds of manufacturing steps and months of production to make chips, the semiconductor industry cannot turn around quickly to meet the rising demand fueled by the pandemic.
After decades of discussing how to work out features as small as a few nanometers in silicon leaves, the spirit of Moore’s Law — the hope that cheap and powerful chips will be available — is threatening something in the world right now. supply chains.
A lonely limit
Twenty years ago, there were 25 manufacturers in the world making cutting-edge chips. Today, alone Taiwan Semiconductor Manufacturing Company (TSMC) Taiwan, Intel In the United States, and Samsung They have facilities or beans that produce the most advanced chips in South Korea. And Intel, a long-time technology leader, is struggling to keep up, as it has repeatedly lost production deadlines in recent generations.
One of the reasons for the consolidation is the construction of a facility to make the most advanced chips cost between $ 5,000 and $ 20 billion. These fab factories make chips with characteristics of a few nanometers; in industry jergan are called 5 nanometers and 7 nanometers nodes. Most of the cost of new Fab factories is spent on buying the latest equipment, such as an an extreme ultraviolet lithography (EUV) machine that costs more than $ 100 million. Made exclusively by ASML in the Netherlands, EUV machines are used to record specific circuit models with nanometer-sized characteristics.
Chip manufacturers have been working on EUV technology for more than two decades. After a billion-dollar investment, EUV machines were first used in the production of commercial chips in 2018. “This tool is 20 years late, 10 times larger than the budget, because it’s amazing,” says David Kanter, executive director of the machine-based open engineering consortium. learning. “It simply came to our notice then. It’s completely like science fiction. ”
Such a major effort made it possible to create millions of tiny transistors on Apple’s M1 chip made by TSMC; It is one of the first generations of cutting-edge chips based entirely on EUV.
Only the largest technology companies are willing to pay hundreds of millions of dollars to design a chip for cutting-edge nodes.
Paying for the best chips makes sense for Apple, as these chips come in the latest MacBook and iPhone models, selling for millions at luxury brand prices. “The only company that is really using high volume EUV is Apple, and they sell $ 1,000 phones for those with a crazy margin,” says Kanter.
In addition to being expensive to manufacture such chips, the cost of designing huge complex circuits is affordable for many companies. Along with Apple, only the largest technology companies that require the highest computing performance, e.g. Qualcomm, AMD, and Nvidia, are willing to pay hundreds of millions of dollars to design a chip for cutting-edge nodes, says Sri Samavedam, vice president of CMOS technologies at an international research institute in Imec, Leuven, Belgium.
Many more companies are producing laptops, televisions and cars that use chips made with old technologies, and the growth in their demand is at the root of the current chip shortage. Simply put, most chip customers can’t afford or don’t want to pay for the latest chips; today a typical car uses dozens of microchips, and an electric vehicle many more. It adds up quickly. Instead, authors like cars have stuck to chips made using older technologies.
What’s more, many of today’s most popular electronics don’t require cutting-edge chips. “It doesn’t make sense, for example, to put up an A14 [iPhone and iPad] we need to chip in every computer we have in the world, ”says Hassan Khan, a former PhD researcher at Carnegie Mellon University who studies the effects of public policy at the end of Moore’s Law and now works at Apple.“ You don’t need your smart thermometer at home, and you don’t need 15 of them in the car because he’s very hungry and very expensive. “
The problem is that as more users rely on older, cheaper chip technologies, the giants in the semiconductor industry are focused on building new cutting-edge fabs. TSMC, Samsung and Intel have recently announced $ 1 billion in investments for the latest manufacturing facilities. Yes, they are expensive, but that’s where the profits lie — and there’s a future for the last 50 years.
TSMC, the world’s largest contract maker, earned nearly 60% of its 2020 revenue by making cutting-edge chips with 16 nanometers and smaller, including an Apple M1 chip made with a 5-nanometer manufacturing process.
The problem is that “no one is building semiconductor manufacturing equipment to support old technologies,” says Dale Ford, of the Electronic Components Industry Association in Alpharetta, Georgia. “So here we are stuck between a rock and a hard place.”
Low-end chips
All of this is important for technology users, not only because of the disruptive supply that is currently emerging, but also because it jeopardizes the development of many potential innovations. In addition to being more difficult to produce, cheaper merchandise chips are also more expensive, as each generation of chips requires more expensive equipment and facilities than previous generations.
Some consumer products will only require stronger chips. The emergence of faster 5G mobile networks and the rise of 5G speed-based computer applications may force investments in specialized chips designed for network equipment that speaks to tens or hundreds of devices connected to the Internet. In automotive functions, such as advanced driver assistance systems and “infotainment” systems, cutting-edge chips can also be used, as demonstrated by the electric vehicle manufacturer. He reported to Tesla to develop a chip with both TSMC and Samsung for future car drivers.
But buying the latest cutting-edge chips or investing in specialized chip designs may not be practical for many companies with “intelligence” in developing future products. Consumers of consumer devices, such as a Wi-Fi-enabled sous vide machine, are unlikely to spend money on developing specialized chips on their own, adding even more stylish features, Kanter says. Instead, they will probably go back to chips that can be made using older technologies.
Most current chip customers come across cheaper raw material chips that represent the link between cost and performance.
And, for example, they say that lower-cost products like clothing have a “fine razor,” which would add a dollar (even less than $ 10 or $ 20) to more expensive labels on the price of each product. This means that rising prices for computing power can prevent the development of clothing, such as the development of voice commands or clothing that can detect and respond to weather changes.
The world can certainly live without sous vide machines, but not having cheaper and more powerful chips would come at a real cost: the end of the era of inventions driven by Moore’s Law and the promise of decades of affordable computing power will be available for future innovation. .
Most current chip customers come across cheaper raw material chips that represent the link between cost and performance. The supply of these resource chips is, in fact, as global demand for computing power grows.
“It’s still happening that the use of semiconductors in vehicles is going up, semiconductors are baking in your toaster and all sorts of things are going up,” says Willy Shih, an internship management professor at Harvard Business School. “So the question is, where will the shortage go next?”
Global concern
In early 2021, President Joe Biden signed an executive order supply chain reviews for the chips and threw his support behind the bilateral push to at least get Congress approved $ 50 billion for semiconductor manufacturing and research. Biden held two White House summits with leaders in the semiconductor and automotive industries, an April 12 meeting in which he highlighted the silicon leaf.
Actions will not soon resolve the imbalance between chip demand and supply. Experts say that, at the very least, the current crisis allows the U.S. government to try to fix the supply chain once and for all and reverse the overall slowdown in semiconductor innovation, and perhaps the U.S.’s ability to make the necessary chips.
Approximately 75% of total chip manufacturing capacity was in East Asia as of 2019, with a U.S. share of about 13%. Taiwan’s TSMC owns nearly 55% of the foundry that handles consumer chip manufacturing orders.
To see it above all is a US-China rivalry. China’s national SMIC champion has been at the forefront of business chip technologies for five or six years. But China’s foundries may help meet global demand for chips built on older nodes in the coming years. “Given the state subsidies they receive, China’s foundries may be the lowest-cost manufacturers in the 22- to 22-nanometer and 14-nanometer nodes,” Khan said. “Chinese factories may not be competitive on the border, but they can supply a larger share of demand.”
[ad_2]
Source link
