Pressure and Heat: How New Technologies Reshape the World

Pressure and Heat: How New Technologies Reshape the World

Export to PDF

Technological Disruption: Lessons from the Rise of Lab-Grown Diamonds

In 1954, a team of scientists at General Electric’s research lab in Schenectady, New York accomplished something that had eluded chemists for over 150 years: they made a diamond. Under a secret initiative called “Project Superpressure,” H. Tracy Hall and his colleagues subjected carbon to 100,000 atmospheres of pressure and temperatures exceeding 2,900°F. Out came tiny, yellowish crystals—too small for jewelry, too rough for engagement rings, but unmistakably diamond. Seventy-one years later, lab-grown diamonds account for over half of all diamond center stones sold in the United States and have crashed the price of a one-carat stone from over $8,000 to under $750. The mined diamond industry is reeling. De Beers is sitting on a $2 billion stockpile of unsold inventory. The disruption is here and it is far from over.

What took so long? And what does any of this have to do with investing in 2026?

The answer to the first question is instructive. Lab-grown diamonds existed for decades before they mattered. GE made gem-quality stones by 1971, yet for the next forty-plus years, they remained a curiosity—expensive to produce, small in scale, and dismissed by an industry that controlled both supply and narrative. The technology simmered. Then, around 2016, two things converged: production costs fell dramatically as Chemical Vapor Deposition (CVD) technology matured, and a generation of consumers arrived who cared more about value and ethics than tradition. The result was a price collapse that industry insiders have dubbed “The Great Correction.” Lab-grown diamond prices have plummeted roughly 96% from their 2018 peak. The tipping point, when it came, was swift and savage.

As Pandora’s CEO Alexander Lacik noted in a Bloomberg interview earlier this year, lab-grown diamonds now represent the majority of all loose stones sold in the United States. Consumers, he observed, “vote with their wallets.” He expects the shift to be permanent. When given the option of identical quality at a fraction of the price, the outcome is inevitable. It just takes a while to get there.

Artificial Intelligence and Knowledge Work: A Similar Disruption Pattern

Now replace “diamond” with “knowledge work” and “lab-grown” with “AI.”

The parallels are striking and instructive. Artificial intelligence as a concept has been around since the 1950s—roughly the same era as those first synthetic diamonds. For decades, AI was an academic curiosity, a novelty that could beat you at chess but couldn’t compose a coherent paragraph. The technology simmered. Deep learning emerged around 2012. Things got more interesting. Then, in November 2022, ChatGPT burst onto the scene and gained 100 million users faster than any internet service in history. Just like lab-grown diamonds, AI went from “curious novelty” to “existential threat” in what felt like a single night. But the seeds had been planted decades ago.

The difference in speed is worth noting. Lab-grown diamonds took 71 years from invention to market dominance. AI is compressing a similar arc into perhaps 15 or 20 years. The pattern is the same—long gestation, sudden disruption—but the clock is running much, much faster.

Early Economic Impacts of AI on the Workforce

We’re still in the early phases. If the AI disruption follows the diamond playbook, we’re somewhere around 2017 in diamond years: the technology works, adoption is accelerating, prices for AI-generated output are plummeting, but the full reshaping of industries is still ahead of us. A recent Harvard Business School study found that since ChatGPT’s launch, job postings for structured, repetitive knowledge tasks—the kind AI handles well—have already declined 13%, while demand for analytical and creative roles has grown 20%. The reshuffling has begun, but it is early innings.

This brings me to my friend—let’s call him Tom. Tom is a senior consultant at one of the large global consulting firms. He’s sharp, experienced, and well-compensated. He’s also convinced he’ll be out of a job in three years. His reasoning is straightforward: a significant portion of what he does—complex valuation scenarios—is precisely the kind of structured knowledge work that AI is learning to do at a fraction of the cost and in a fraction of the time. Tom isn’t panicking. He’s adapting. He’s enrolled in welding classes at his local junior college.

Let that sink in. A well-paid consultant at a marquee firm is learning to weld. Not because he loves metalwork (though he might), but because he sees the writing on the wall: AI can’t weld a pipe. AI can’t wire a house. AI can’t fix a boiler. The physical world remains stubbornly resistant to automation, and the people who work in it are becoming more valuable, not less.

The Great Divergence: AI, Skilled Trades, and Labor Market Shifts

This observation leads to what I’ll call The Great Divergence. AI is exerting deflationary pressure on knowledge workers and inflationary pressure on physical workers. The economics are simple. On the knowledge side, if an AI tool can do in ten minutes what used to take a junior analyst ten hours, the demand for junior analysts drops. More supply of output, less demand for the humans who produce it. Prices—in this case, wages—fall. On the physical side, we have the opposite problem. Baby boomers are retiring in droves, younger generations were steered toward college degrees instead of trades, and the infrastructure demands of the modern economy—data centers, EV charging networks, grid expansion, renewable energy installations—are exploding. The Bureau of Labor Statistics projects 81,000 annual openings for electricians alone through 2034, and there aren’t enough qualified workers to fill them. Fewer workers chasing more demand means wages go up. Way up.

Here’s the irony that would have been unthinkable a generation ago: the kid who skipped college to become an electrician may end up earning more than the kid who got an MBA and went into consulting. Specialized welders already command six figures. Master electricians running their own shops exceed $100,000. Meanwhile, PwC’s 2026 AI predictions describe the knowledge workforce reshaping into an “hourglass”—plenty of room at the very top for senior strategists and at the very bottom for AI-savvy junior workers, but a hollowed-out middle where mid-career professionals like Tom find themselves squeezed.

Investment Implications of AI and Workforce Transformation

Is Tom right to be worried? Probably. Is he right to learn welding? We think so—and not just as a fallback. The skilled trades represent one of the few areas where supply-demand dynamics appear to be moving decisively in workers’ favor. A CSIS study from last year highlighted that even the most conservative AI infrastructure buildout scenario requires over 63,000 additional skilled tradespeople beyond normal growth—and the aggressive scenario pushes that number above 140,000. These aren’t jobs that can be trained overnight. Electricians need four to five years of apprenticeship. Welders need certification and years of practice. The pipeline is thin and the demand is enormous.

For investors, this divergence creates identifiable opportunities. On the deflationary side, companies that successfully deploy AI to replace expensive knowledge work could see margin expansion. The winners may be firms with clearly defined AI strategies that deliver measurable cost savings—not the ones buying AI tools for the sake of a press release. On the inflationary side, companies serving the skilled trades—industrial equipment manufacturers, trade education providers, workforce management platforms, and the electrical and infrastructure supply chain—could stand to benefit from years of pricing power driven by labor scarcity.

The Long Arc of Technological Disruption for Investors

The diamond lesson is a patient one. Disruptions take much longer to arrive than the headlines suggest, but when they hit, they hit hard and they don’t reverse. Nobody is going back to paying $8,000 for a lab-grown diamond. And nobody is going back to a world where it takes ten analysts to produce a market report that AI can draft in minutes. The question for investors isn’t whether these disruptions are coming—they’re already here. The question is whether your portfolio is positioned on the right side of the pressure.

It took 100,000 atmospheres of pressure and 2,900 degrees of heat to make the first synthetic diamond. Artificial intelligence is applying similar forces to the labor market. Some things will crack. Some things will be forged stronger. Knowing the difference is what we strive to do.