Magnificent Seven: Will AI Make You Rich, or Are You Buying the Decade's Most Expensive Bubble?

Maybe you heard about it on a podcast, or it popped up on YouTube. Magnificent Seven — seven companies that dominate global markets and have split investors into two camps.

One side claims we're witnessing the greatest technological revolution since the invention of the internet. The other warns that this whole thing smells like the bubble of 2000.

Who's right? Let's take a look.

What Are the Magnificent Seven?

The name "Magnificent Seven" was coined by Bank of America strategist Michael Hartnett in May 2023. It references the classic western film and refers to seven tech giants:

Company	What They Do	Market Cap (December 2025)
Apple	iPhones, ecosystem	~$4.1 trillion
Microsoft	Cloud, Office, AI	~$3.7 trillion
Nvidia	AI chips	~$4.5 trillion
Amazon	E-commerce, AWS cloud	~$2.4 trillion
Alphabet (Google)	Search, advertising, cloud	~$3.8 trillion
Meta (Facebook)	Social networks, advertising	~$1.7 trillion
Tesla	Electric vehicles, (non-)autonomous driving	~$1.5 trillion

Together, these companies are worth over $21 trillion.

For perspective — that's more than the entire stock markets of Japan, the UK, and France combined.

Why Is Everyone Talking About Them?

Here's where it gets interesting.

These seven companies make up 35% of the S&P 500 index. That means if you buy an S&P 500 ETF (like the popular VOO), more than a third of your money goes into these seven stocks.

Magnificent Seven weight in S&P 500

And their performance? From 2015 to 2024, they returned 697% — while the rest of the S&P 500 returned only 178%.

But here's the thing...

The more something grows, the more people ask: "Isn't this too much?"

And that's exactly where the AI bubble debate begins.

What Is a Bubble, Anyway?

Before we go further, let's clarify what a bubble means.

A bubble occurs when the price of an asset (stocks, real estate, tulips...) rises significantly above its actual value. People buy not because they believe in long-term value, but because they believe the price will keep rising and they can sell for more later.

Classic bubble symptoms:

"This time it's different" — new paradigm thinking
Extreme optimism and FOMO
Prices disconnected from profits and revenues
Massive debt-fueled speculation

Sound familiar? Let's look at both sides of the argument.

Arguments FOR the Bubble: "This Will End Badly"

1. Valuations Are Historically Extreme

Some AI companies trade at absolutely insane multiples.

Palantir? 600x annual earnings. Tesla? 190x annual earnings.

For comparison — the historical average for the S&P 500 is around 15-20x.

OpenAI reached a valuation of $500 billion in 2025 — yet the company isn't profitable and admits it won't be until 2029.

2. AI Investment vs. AI Revenue

This might be the skeptics' strongest argument.

Companies are investing an estimated $400 billion in AI infrastructure this year. But actual revenue from AI services? Maybe $15-20 billion.

A 2025 MIT study found that 95% of organizations haven't achieved any return on investment from generative AI.

That's a massive gap.

3. "Circular Financing"

This reminds me of 2000.

Nvidia invests in OpenAI. OpenAI buys chips from Nvidia. Microsoft owns a stake in OpenAI and is also a customer of CoreWeave, where Nvidia has a stake.

Money flows in circles between the same players. It looks like growth, but it's partly an optical illusion.

4. What the Skeptics Say

"We're definitely in a bubble — probably at 80% of the euphoria we saw in 1929 and 2000." — Ray Dalio, founder of Bridgewater Associates

"Someone is going to lose a phenomenal amount of money." — Sam Altman, CEO of OpenAI (yes, that Sam Altman)

Jamie Dimon, CEO of JPMorgan, sees a 30% probability of a significant correction.

Arguments AGAINST the Bubble: "This Time It Really Is Different"

1. These Companies Are Brutally Profitable

And here's the crucial difference from 2000.

The Magnificent Seven earned over $131 billion in net profit in Q4 2024 alone. Their average profit margin is 25.8% — nearly double the S&P 500 average.

This isn't Pets.com, which sold dog food at a loss of 26 cents on every dollar of revenue.

2. AI Revenue Is Real (At Least for Some)

Nvidia — Data center revenue: $51 billion per quarter (+66% year-over-year). Profit margin: 73%. Chips are sold out through 2026. This is real money for real hardware.

Microsoft — Azure AI adds 12 percentage points to cloud growth. GitHub Copilot has 1.3 million paying subscribers. They're heading toward $10 billion per year in AI revenue.

Google/Alphabet — Cloud business: $43 billion annually, growing 30%. Backlog: $155 billion. Gemini has 650 million monthly active users.

3. Valuations Aren't as Crazy as 2000

Here are the numbers:

Metric	Dot-com 2000	AI 2025
P/E of top 5 companies	50x	28x
P/E of tech leaders	59x	34x
Nasdaq P/E	200x	~35x

Goldman Sachs summarizes it:

"Valuations are stretched, but not yet at levels consistent with historical bubbles."

4. What the Optimists Say

Jerome Powell, Fed Chair, in October 2025:

"These companies have real profits... business models and earnings. This is truly something different from the dot-com bubble."

Cathie Wood believes Tesla will reach $2,600 per share within five years. Jensen Huang of Nvidia predicts that global data center investments will grow from $600 billion to $3-4 trillion by 2030.

A Brief History Lesson: The Dot-com Bubble of 2000

To understand today's situation, you need to know the history.

What Happened?

In the 90s, everyone believed the internet would change the world. And they were right — the internet really did change the world.

But stock prices grew much faster than reality. Companies with no revenue, no profits, often no real product achieved billion-dollar valuations.

Then it burst.

Dot-com bubble — Nasdaq

Nasdaq fell 78% over 31 months
It took 15 years to return to its previous highs
85% of dot-com companies went bankrupt by 2004
Pets.com burned through $300 million in 268 days
Webvan lost $1.5 billion

What Did It Mean for Regular Investors?

A Vanguard study found that 70% of 401(k) accounts lost at least a fifth of their value.

People invested $260 billion into stock funds in 2000 — buying at the absolute peak. Meanwhile, insiders sold $43 billion in shares.

Similarities to Today

Massive capital concentration (58% of VC investments go to AI)
"New paradigm" thinking
Extreme market concentration
Infrastructure building with unclear returns

Differences from Today

Dot-com 2000	AI 2025
Companies lost millions quarterly	Companies earn billions
Pets.com: -26% margin	Nvidia: +55% margin
No real products	Real products in production
P/E 70x (forward)	P/E 26x (forward)

How Do the Magnificent Seven Actually Make Money from AI?

This is the key question. Let's go through each company.

Nvidia — Clear Winner

Data center revenue: $51 billion per quarter (+66% year-over-year). Profit margin: 73-75%. AI accelerator market share: 80-90%.

These are real chips, really sold, really deployed. Nvidia is a clear case of genuine AI revenue.

Microsoft — Solid

Azure AI adds 12 points to cloud growth. GitHub Copilot: $2 billion annual revenue. Microsoft 365 Copilot growing. Heading toward $10 billion per year from AI inference.

Google/Alphabet — Growing

Cloud: $43 billion annually, +30%. Vertex AI customers grew 5x. Backlog: $155 billion.

Amazon AWS — Solid

"Multi-billion dollar annual run rate, growing triple digits." Bedrock, SageMaker, Trainium2 chips in production.

Meta — Indirect

AI improves ad targeting → higher CPM → higher revenue. Meta AI: 700 million monthly users. Llama models: 650 million downloads.

But note — Meta hasn't directly monetized AI yet.

Apple — Weak

Apple Intelligence is free with every device. No direct AI revenue. Siri upgrade delayed. 1-2 years behind competition.

Tesla — Mostly Promises

Full Self-Driving: ~$1 billion annually (deferred revenue recognition). Robotaxi: no commercial revenue. Optimus robot: "just a few hundred units" by mid-2025.

Elon Musk claims Optimus will represent "80% of Tesla's future value." But for now, it's just presentations and promises.

What Should You Take Away?

1. The Truth Is Somewhere in the Middle

This isn't a pure bubble like 2000 — these companies are brutally profitable. But it's not a "safe bet" either — valuations are historically stretched.

2. Not All AI Stocks Are Equal

Nvidia sells real chips for real money. Tesla sells promises about robotaxis.

That's a huge difference.

3. Historical Lesson

Even revolutionary technologies can be bad investments if you buy them at the wrong price.

The internet really did change the world. But anyone who bought the Nasdaq in March 2000 waited 15 years to get their money back.

4. For the Global Investor

If you invest in global ETFs (VWCE, IWDA), you already have the Magnificent Seven automatically. They make up a large portion of your portfolio whether you want them or not.

That's not necessarily bad — but it's good to know what you're investing in.

Diversification remains your best friend.

Conclusion

Are we in an AI bubble? Honestly — nobody knows for sure. Everyone has an opinion, and everyone's acting like they know everything.

Personally, I continue to choose global ETFs as the majority of my portfolio, and for the more fun "stock picking" portion, I look at companies that provide AI infrastructure (like Amazon with AWS). Because I work with various AI tools daily and I see that plenty of them pop up and disappear every week, and I don't see a clear winner yet, except for those who provide them with data centers...

Have questions or a different perspective? Reach out at vymerd@gmail.com.