Listen to the audio version of this article (generated by AI).
Editor’s Note: On Tuesday, we looked at how AI is making the oil and gas industry more efficient with Joe Austin. Today, we’re looking at another place where AI is making a big difference: semiconductor manufacturing.
A single defect in a semiconductor can cost as much as $25,000, and Joe explains why AI is becoming the only practical way to catch those problems before they become expensive mistakes.
It’s another reminder that some of AI’s biggest opportunities may not come from the companies building the technology. They may come from the businesses using it to become more efficient and more profitable.
That’s one reason Marc Chaikin of Chaikin Analytics recently introduced his new AI-Powered Time Machine. It’s designed to help investors look beyond today’s biggest AI stocks and uncover the next generation of potential winners.
Yesterday, Marc and Joe unveiled the AI-Powered Time Machine for the first time. If you missed the presentation,
Now, here’s Joe with a closer look at how AI is transforming manufacturing one factory at a time…
California Steel Industries’ Hot Strip Mill in Fontana stretches more than half a mile long.
Inside, giant ovens heat steel slabs to about 2,300 degrees Fahrenheit. At that temperature, the steel gets soft enough to roll.
But first, it needs cleaning. The furnace leaves a thick crust of “scale” on the surface. If it isn’t removed, it gets pressed into the steel and ruins the finish. A scalebreaker cracks the crust loose. Then high-pressure water jets blast it away.
Next, the steel slab passes through five roughing stands that squeeze it down from between 7 and 9 inches thick to as little as 0.0538 inches — close to the thickness of a credit card. A crop shear trims the ragged ends before the steel moves to finishing. Then six more finishing stands roll it to its final thickness and surface quality.
By this point, the steel is moving at about 35 miles per hour.
That’s too fast to catch defects by eye. For automotive panels and appliances, the surface has to be flawless — defects show right through the paint.
The finished strip winds into a coil. Some weigh up to 25 tons. The whole process takes about five hours. At full capacity, the mill runs 24 hours a day and produces 2 million tons of steel per year.
But at least you can see steel.
In today’s most advanced semiconductor fabrication plants, the defects that matter are invisible to the human eye.
And the consequences of missing them are just as severe.
A Single Semiconductor Defect Can Cost $25,000
In semiconductor manufacturing, everything starts with a wafer — a thin, polished disc sliced from pure silicon, usually about 12 inches across. These wafers must be flawless. Even a microscopic scratch or contaminant can create defects across hundreds of chips.
The first step is circuit printing using extreme ultraviolet lithography. This process projects circuit patterns using light with a wavelength shorter than any visible color. A single finished chip can require 20 to 30 passes through this stage alone.
The specialized masks used in this process — a kind of three-dimensional stencil — have to be perfect, too. A single defect ruins every chip that mask touches. And those masks can cost up to $1 million each.
After each pass, the wafer goes through etching, deposition, and chemical treatment to build up transistor layers. Then the cycle repeats. Today’s most complex chips go through 1,500 to 2,000 individual steps before they become functional. Each step is a potential failure point. One particle of dust can ruin an entire wafer.
A single wafer for the most advanced semiconductors can cost between $20,000 and $25,000. Each wafer holds hundreds of chips. A defective one wipes out hundreds of products at once. And the fabs where all this happens cost between $15 billion and $20 billion to build.
Fabs need to reduce these losses wherever possible. And human inspectors simply can’t do the job. At 35 miles per hour, steel moves too fast to see. In a semiconductor fab, the defects are too small to see. In both cases, the stakes are too high to miss anything.
AI Is Boosting Quality Control
This is one area where AI doesn’t just help. It’s the only solution that actually works.
AI “deep learning” and “edge learning” take defect control to a level humans can’t match. Deep learning works by analyzing hundreds of example images until the system learns to make decisions on its own — no programmer required at each step.
Edge learning goes further. These systems come pretrained and may need as few as five to 10 images to get started. They deploy in minutes.
The results are measurable.
At BMW, AI-powered vision systems cut defect rates by 30% at one European plant within a year. Customer satisfaction jumped 15% after the rollout. At Foxconn, AI-powered cameras now catch defects with 98% accuracy, flag 80% fewer false alarms, and inspect each unit 60% faster than before.
These aren’t pilot programs. They’re production systems running at scale, in some of the most demanding manufacturing environments on Earth.
This is what I mean when I say the real AI story isn’t the one getting the most attention.
Everyone is watching the big infrastructure names — the chip companies, the cloud providers, the chatbot platforms. And yes, those are important. But there’s a parallel story playing out on the factory floor, in the oil field, and in the semiconductor fab.
AI is solving problems that weren’t solvable before. And the companies delivering those solutions are becoming more competitive, more profitable, and more valuable — quietly, without much fanfare.
That’s exactly the kind of opportunity I’ve spent my career looking for.
Finding the Next Generation of Winners
The challenge, of course, is identifying which companies are actually winning — not just claiming to use AI, but using it in ways that show up in the fundamentals.
That’s a problem Marc Chaikin has been working on his entire career. His Power Gauge rating system was built to cut through the noise and find stocks with real momentum behind them. It’s been doing that for decades.
Marc and I are going a step further. We unveiled the first AI-powered product Chaikin Analytics has ever built — and it’s unlike anything we’ve shown the public before.
We’re calling it the Time Machine. It scans decades of market history to find stocks today whose fundamental and technical fingerprints match the early profiles of stocks like Nvidia Corp. (NVDA), Amazon.com Inc. (AMZN), and Meta Platforms Inc. (META) — right before they made their biggest moves. In backtesting, it surfaced stocks that went on to deliver gains of 995%, 1,406%, and 3,804%, all while the “seed” stocks they were matched against posted far more modest returns.
The factory-floor AI story is one example of the kinds of opportunities the Time Machine is designed to surface. Companies solving real industrial problems with AI — before the market catches on.
This is the first time we’ve ever made something like this available to individual investors.
Good investing,
Joe Austin
Senior Analyst, Chaikin Analytics
P.S. Most investors are focused on the AI names everyone already knows. Joe is focused on the ones most people haven’t found yet — the companies using AI to solve problems in places like the factory floor and the oil field, before Wall Street fully catches on. He and Marc Chaikin debuted the first AI-powered tool Chaikin Analytics has ever built to help find exactly those stocks.
