Why Is AI Investment Not Creating Jobs Like the Internet Boom Did?

Mar 24, 2026

The Problem: Where Are All the Jobs?

I keep hearing the same question: AI companies are worth hundreds of billions. NVIDIA makes over $100 billion a year. OpenAI is valued at $300 billion. So where are all the jobs?

When the internet boom happened, companies hired thousands. When mobile took off, millions found work as drivers, delivery riders, and content creators. But AI? The numbers tell a different story.

Let me break down why this is happening.

The Internet Era: Jobs at Every Layer

When I look back at the internet and mobile waves, I see money flowing to people.

Investment → Salaries → Mass Hiring
                    ↓
            Product Managers
            Developers
            Designers
            Operations Staff
            Customer Support
            Marketing Teams
            Content Editors
            Delivery Riders
            Drivers
            Content Creators

Think about Meituan at its peak: 60,000 to 70,000 employees. ByteDance? Over 100,000 people. A single food delivery app didn’t just employ office workers—it supported millions of riders.

Mobile platforms created entire ecosystems:

Platform Type	Jobs Created	Example Scale
Food Delivery	Millions of riders	Meituan, Ele.me
Ride-hailing	Millions of drivers	Didi, Uber
Live Streaming	Tens of millions of creators	Douyin, TikTok
E-commerce	Millions of sellers	Taobao, JD

The internet required humans at every step. Someone to write code. Someone to design interfaces. Someone to handle support. Someone to manage operations.

The AI Era: Money Flows to Hardware

Now let me show you what happens with AI investment.

Investment → Hardware → Minimal Human Touch
                    ↓
            GPUs (NVIDIA)
            Data Centers
            Cloud Computing
            Electricity

When a company spends $100 million on AI, where does it go?

First, chips. They buy GPUs from NVIDIA. NVIDIA contracts TSMC for manufacturing. TSMC’s production lines? Highly automated. The semiconductor supply chain creates very few jobs per dollar spent.

Second, compute. Training a large language model costs tens or hundreds of millions. That money goes to AWS, Azure, or GCP. These providers build data centers, buy servers, pay for electricity. Data centers need workers—but mostly for operations, power management, and security. Not high-paying technical roles.

Third, a tiny elite. AI companies need very few people, but those people must be exceptional.

The Headcount Reality Check

Let me compare headcounts directly.

Company Type	Era	Valuation (Approx)	Employees
Meituan	Internet	$100B+ peak	60,000-70,000
ByteDance	Internet/Mobile	$200B+	100,000+
OpenAI	AI	$300B	~2,000
Anthropic	AI	$40B+	<1,000

See the difference? AI companies at similar valuations have one to two orders of magnitude fewer employees.

This isn’t a bug—it’s the feature. AI’s core value proposition is doing more with fewer people.

Why the Difference Matters

I think this shift has real implications for different groups.

For job seekers: The old path of “learn to code, get a job” doesn’t scale the same way. The internet era needed thousands of average developers. AI needs dozens of exceptional ones.

For investors: Returns concentrate among fewer winners. Less spillover to broader employment means narrower wealth distribution.

For policymakers: Traditional strategies—like training programs for tech workers—may not produce the same results.

For educators: Preparing students for “average” tech roles faces diminishing returns.

Common Misconceptions I Hear

”AI Will Create New Job Types”

This argument worked for previous tech waves. Cars created drivers, mechanics, gas station workers. The internet created SEO specialists, content moderators, social media managers.

But here’s the key difference: those jobs emerged because the technology required human labor at scale. AI’s promise is precisely the opposite—reducing human labor, not creating new demands for it.

”Application-Layer AI Will Need Developers”

I’ve seen this up close. Building on AI today often means:

Calling an API
Writing wrapper code
Iterating on prompts

What previously required 5-6 people (product manager, designer, frontend dev, backend dev, tester, operations) can now be done by 1-2 developers using tools like Cursor or Claude Code.

The application layer is thin. The value sits in the model, not the wrapper.

A Simple Mental Model

I find this comparison helpful:

Internet Era:
  ├── Salaries: $60M (hundreds of employees)
  ├── Marketing: $20M
  ├── Infrastructure: $15M
  └── Other: $5M

AI Era:
  ├── Compute/Cloud: $70M
  ├── Hardware/GPUs: $20M
  ├── Salaries: $8M (dozens of elite staff)
  └── Other: $2M

Same investment. Very different job creation.

Summary

In this post, I explored why AI investment isn’t creating jobs like the internet boom. The core reason is simple: AI money flows to chips and compute, not salaries. AI companies operate with tiny, elite teams—a few hundred or thousand people instead of tens of thousands. This isn’t an accident or a temporary state. It’s fundamental to how AI creates value.

The implications matter for anyone counting on tech to drive employment growth. The old formulas—more investment equals more jobs—don’t apply the same way. AI concentrates wealth among hardware suppliers and a small technical elite. The few AI jobs that exist require exceptional qualifications, not the mass-employable skills of previous eras.

Understanding this shift helps us have more realistic expectations about AI’s economic impact and plan accordingly.

Final Words + More Resources

My intention with this article was to help others share my knowledge and experience. If you want to contact me, you can contact by email: Email me

Here are also the most important links from this article along with some further resources that will help you in this scope:

👨‍💻 Reddit Discussion: AI Job Creation

Oh, and if you found these resources useful, don’t forget to support me by starring the repo on GitHub!