AI, Unemployment, and the Solution to Future Job Loss

I hesitate to write this article because it risks sounding alarmist. But as an investor, a husband, a father, and a wage earner, I feel obligated to look at the world as it is, not as I wish it to be. Hope is not a hedge. Optimism is not a plan. And when structural forces begin reshaping the economy at a pace we haven’t seen in decades, ignoring them is not prudence, it’s negligence.

It’s in that spirit that I approach the topic of systemic, AI‑driven unemployment—a conversation that has been accelerating as of late, especially among those of us in IT who sit on the front lines, and among economists and market practitioners who are watching the data shift in real time.

The piece that crystallized this issue for me wasn’t a white paper from a major bank or a think tank. It was a short article by retail investor Leo Nelissen aptly titled "The Old Playbook Is Dead - And Wall Street Has To Adapt". The focus of the article is on the jobs-to-GDP relationship. Nelissen highlights a troubling divergence: GDP growth has been trending upward since 2023, while nonfarm payroll growth declines. If the trend continues, we may soon see a historically unusual scenario—employment falling while GDP rises.

Credit: Nelissen

That divergence is not normal, but it does tell a story. It aligns disturbingly well with the idea that AI is beginning to eat the world. At present this threat is most prominent in IT, where coding and software are the first targets.

Earlier this month, the market experienced what many have called a SaaS apocalypse - the “SaaSpocalypse.” Between mid‑January and mid‑February 2026, roughly $1 trillion in software market value evaporated as investors reassessed the future of traditional SaaS firms.

Why? Because SaaS companies are built on armies of engineers, salespeople, and support staff—precisely the roles AI is now automating. New AI tools from Anthropic and others triggered a broad sell‑off after demonstrating that AI agents can perform, accelerate, or outright replace many of the tasks that underly SaaS business models and revenue streams.

This is not a normal cyclical correction. It is a repricing of labor‑intensive business models.

And that brings us to the macro picture.

AI Capex: The New Engine of Growth

The divergence between GDP and employment becomes easier to understand when you look at where growth is coming from. AI‑related capital expenditures - data centers, chips, model training, and automation infrastructure - have become a meaningful driver of GDP growth. Economist Jason Furman estimates that AI capex accounted for 92% of GDP growth in early 2025 despite being only 4% of GDP.(1)  Notably, contributions this size are outpacing consumer spending and upending the traditional consumer-driven economy.

This is the definition of a structural shift, and unlike traditional capex, which often creates jobs, AI capex replaces them.

Companies are pouring billions into compute because it offers an insanely attractive alternative to labor. AI never sleeps, doesn’t get sick, doesn’t require health care or a 401(k), and works for a fraction of the cost.

When you combine rising AI investment with falling labor demand, the macro picture begins to make sense: GDP can rise even as employment falls—because capital, not labor, is doing the work.

When Capital Learns to Build Itself

This shift isn’t just visible in economic data—it’s visible in the technology itself. In November, 2025 Anthropic announced Claude Opus 4.5 as a major step forward and highlighted that the model was developed with significant assistance from earlier Claude systems - effectively AI improving AI. And a trend of self-improving capital is has profound implications for the value of human labor.

“Something Big Is Happening”

Matt Shumer’s widely circulated essay Something Big Is Happening captures the speed and magnitude of the shift. He argues that we are witnessing a technological inflection point where AI systems are rapidly becoming capable of performing a broad range of cognitive tasks once reserved for humans.

I encourage everyone to read his essay for themselves, but there are two striking observations among others he raises. First, AI is improving faster than anyone expected, compressing years of capability gains into months. Second, the number of tasks AI can perform at or above human level is expanding exponentially, not linearly. We have reached the knee of the technological improvement curve. New advances are built on top of yesterday's at an ever increasing pace. The aforementioned example with Opus is a prime example, and there is more to come. 

Shumer’s tone is not apocalyptic, but it is urgent. He frames the moment as one where knowledge workers must adapt quickly or risk being left behind. What's more there is unlikely to be any relief in the near term. With AI accelerating this quickly, the labor market will feel the pressure long before policymakers or institutions can react.

Ghost GDP

If Shumer describes the present, Citrini Research sketches a possible future. Their piece titled 2028 Global Intelligence Crisis is speculative, but it paints a possible vision of what an AI‑dominated economy might look like. They describe a world where AI becomes the primary driver of corporate efficiency, Human labor - especially expensive, white‑collar labor - is systematically displaced, GDP continues to rise even as employment stagnates or declines, creating what they call “Ghost GDP” - growth unaccompanied by broad‑based prosperity.

This is not a forecast. It is a scenario. But it is a scenario that aligns disturbingly well with the data we are already seeing and the capabilities Shumer describes. The message?

AI is reshaping the economy faster than labor markets can adapt.

And that leads to the next question: what kind of economy does this create?

The K‑Shaped Economy: The New Divide

If these trends continue, the future will be K‑shaped. A K‑shaped economy is one where those with capital, skills, or ownership of AI tools rise sharply while those dependent on wages—especially routine cognitive wages—fall (hence the top and bottom, diverging legs of the letter K.). The middle class, which historically relied on stable white‑collar employment, is at risk of being hollowed out. The dividing line between the upward and downward branches of the “K” is simple:

Capital.

Not talent. Not work ethic. Not education. Capital.

Those who own productive assets - equities, real estate, intellectual property, or AI systems - will benefit from the compounding effects of automation. Those who rely solely on labor income will face increasing pressure as the value of human labor declines relative to machine labor.

This is the uncomfortable reality many are reluctant to acknowledge. But acknowledging it is the first step toward preparing for it.

A Lesson From Scripture

In the book of Genesis, Joseph interprets Pharaoh’s dream: seven years of abundance will be followed by seven years of famine. Joseph’s counsel is simple - store up during the good years so the nation can survive the lean ones. Egypt listened. They prepared. And when the famine came, they endured while others perished.

We are living in the “years of plenty” right now—high employment, strong GDP, rising markets. But beneath the surface, the structure is shifting. The labor market is softening. AI is accelerating. And the window to prepare is narrowing.

But it's not too late.

Get Your Financial House in Order

Shumer offers several pieces of advice for navigating this transition, but one stands above the rest as a Joseph-like directive: Get your financial house in order. This means:

• If you have debt, pay it down.

• Avoid large purchases that assume your current income will continue indefinitely.

• Reduce fixed obligations.

• Build liquidity.

• And once you’ve done that: invest. Not recklessly. Not speculatively. But deliberately.

Define your investment horizon. Understand your risk tolerance. Clarify your savings goals. Construct a portfolio that aligns with the world we are actually moving into—not the world we grew up in.

For many of us in the aging workforce, putting our hard earned capital to work in the right invest vehicle is perhaps the most impactful thing that can benefit those most affected by this - our children. And as AI is eating the world rapidly, there no time left to sit on the sidelines. 

(1) Godwin, Paul Ugbede, "AI Data Centers Contributions to U.S. GDP Growth Outpace U.S. Consumer Spending", Tekedia, November 23, 2025, https://www.tekedia.com/ai-data-centers-contributions-to-gdp-growth-outpace-u-s-consumer-spending/