Computer Chips Are the Most Magnificent Things Humanity Has Ever Made

Author: Gideon Lewis-Kraus, The New Yorker (Annals of Technology, issue of October 26, 2026; online October 19, 2026) URL: https://www.newyorker.com/magazine/2026/10/26/computer-chips-are-the-most-magnificent-things-weve-ever-made Saved via: Patrick McGee’s tweet (October 23, 2026): “Such a great article — not necessarily full of new info if you follow chips, but it’s wonderfully written.” One-line: The semiconductor supply chain, not the moon landing, is modern civilization’s monument. It was built by intercontinental cooperation and tacit craft knowledge rather than by lone genius or market competition, which is why no country can simply “just” rebuild it at home.

A long reported feature (about 8,800 words) built on visits to ASML in Veldhoven and Connecticut and to TSMC’s Hsinchu headquarters and AP6/AP7 advanced-packaging fabs, plus interviews with Pat Gelsinger, Dylan Patel, Audrey Tang, and Mike Schmidt (who ran the Biden Administration’s CHIPS Act implementation). As McGee said, the history will be familiar from Chip War. The reporting from inside the fabs and the closing argument are what’s new.

Key claims

• The supply chain is the monument. The piece opens by ranking it alongside Machu Picchu, the Great Wall, the Pyramids and the Roman aqueducts: “Modern global civilization has made one contribution to this eternal pantheon. It is not the moon landing. It is the semiconductor supply chain.” A chip is “a functional city for a population of busy electrons,” engraved at atomic scale and stamped out by the billion.
• The stakes are economic and geopolitical. Nvidia is the world’s most valuable company and its stock delivered about fifteen cents of every dollar the S&P 500 returned last year. A Taiwan war’s first year could cost $10.6 trillion, about a tenth of global GDP (Bloomberg Economics). Gelsinger: Taiwan is “eighty per cent of the supply chain of the world,” and a three-week brownout there would outdo the Great Depression. Chinese chipmakers are about five years behind.
• The history: government as first customer, then commercial scale. From the Bell Labs transistor through Kilby (Texas Instruments) and Noyce (Fairchild, then Intel), the first integrated circuits were affordable only to the federal government. Defense contracts built the manufacturing base, and commercial demand drove scale. Gordon Moore’s 1964 extrapolation (Moore’s Law) became a self-fulfilling promise kept by more than two million industry employees.
• EUV was a bet on the nearly impossible. By the early 1990s deep ultraviolet looked like the end of the road, around 2005. Intel and the U.S. government backed extreme ultraviolet, and only ASML, working with Zeiss, could build it. Samsung’s first order in 2006 needed 100 watts of light when ASML was stuck at five, and ASML missed the deadline by seven years. Today a machine the size of two stacked fire trucks ships in seven 747s, fires molten tin 60,000 times a second to make plasma fifty times hotter than the sun’s surface, and ASML makes only about fifty a year.
• There is no blueprint. The intellectual property “is spread out across the entire system,” across suppliers of suppliers (the glass recipe belongs to Zeiss’s supplier). Parts have two-year lead times, and some are deburred by hand at family shops. The industry “better resembles a kind of coöperative, or a planned economy” than a textbook market.
• Morris Chang and the learning curve. Chang, passed over at Texas Instruments, was recruited by Taiwan’s government and built the foundry model that Gordon Moore told him wasn’t a good idea, against Jerry Sanders’s “real men have fabs.” TSMC rescued a near-bankrupt Nvidia, has made every iPhone chip since 2014, and runs on vigilance: a 209-second idle on one polishing tool is a topic at the shift handover. One in forty-six babies in Taiwan is born to a TSMC employee.
• Advanced packaging is the new bottleneck. AP6 stitches chips into composite superchips for “only our top five customers.” VP Jun He has had to roughly double capacity every year since AI demand hit: “Our key comparative advantage is the management of complexity.”
• Why Arizona is slow. TSMC’s Phoenix fab took about four years versus eighteen months for a Taiwanese phase. The contractor pipeline didn’t transfer, half the initial staff came from Taiwan, and Phoenix has no clean-room laundry good enough, so bunny suits are shipped back and forth. A chip crosses borders more than seventy times. Most steps are monopolies or duopolies, from eleven-nines polysilicon to the North Carolina quartz deposit. Schmidt calls the CHIPS Act “quite self-consciously a generational program.”
• The closing argument is about cooperation. Every confident AI prediction assumes ever-better chips, and “Moore’s Law has survived only on the basis of unprecedented international confederation.” Lewis-Kraus ties this to Henry Adams’s Law of Acceleration and the Dynamo, and to Stonehenge, which was not a lost-technology mystery but many people hauling stone together.

Notable quotes

There are a handful of truly monumental collective achievements in our history: Machu Picchu, the Great Wall of China, the Pyramids of Giza, and perhaps, if we’re being generous, the Roman aqueducts. Modern global civilization has made one contribution to this eternal pantheon. It is not the moon landing. It is the semiconductor supply chain.

The Apollo mission’s chips were less powerful than those currently inside a digital pregnancy test. We squander moon-landing capability to display a plus sign.

“There simply is no ‘file’ of our machines. The intellectual property is spread out across the entire system.” (Marc Assinck, ASML)

When you dig down to the very bedrock of A.I., you discover that underneath all of its cognitive power is a dude drumming his fingers with occult proficiency.

The semiconductor supply chain is a planetary superorganism with its own ecology. It’s not like a plant that can be casually repotted.

“Rockets are an amazing accomplishment, yes. We don’t have a rating scale for complexity, but rockets are maybe a five or a six.” … “Semiconductors are an eleven.” (Brinton Johns and Jon Bathgate, N.Z.S. Capital)

The chief determinant of that unlikely success has been not individual genius in the lab, or individual prowess in the boardroom, but collective planning and coördination at a breathtaking scale.

How it connects

• Chips For America: Kyle’s essay on domestic chip manufacturing. The Arizona section is the on-the-ground version of that question, including Schmidt’s “in order to do it you have to do it” case for the CHIPS Act.
• Chip War: Chris Miller’s book is quoted directly (“Chang wanted to become the Gutenberg of the digital era. He ended up vastly more powerful.”). The Thinking Machine (Stephen Witt) supplies the Jensen Huang / Morris Chang phone-call story, and The Chip (T. R. Reid) the calculator history.
• Semiconductors, TSMC, ASML, Intel, Nvidia, Taiwan, Moore’s Law: the entities the piece is about.
• Gordon Moore, Jensen Huang, Chris Miller, Dylan Patel: people quoted or profiled.
• The Education of Henry Adams: the source of the Dynamo and Law of Acceleration framing in the ending.
• The End of Startups: the demand-side view of chip scarcity, where compute access is rationed by credit.