💢💢💢

🔥🎯 The Top Performers of the S&P 500 Over the Past 10 Years: The Real Winners Aren't AI, but the "Underlying Infrastructure"

If you only look at the top spot, it's easy to draw a simple conclusion:

AI won.

$NVIDIA(NVDA.US) rose +19,528% in ten years

But when you look at the entire list, what becomes truly clear isn't "AI," but another, more fundamental thread:

👉 Who is supporting AI

The full list itself already provides the answer:

$NVIDIA(NVDA.US)
+19,528%

$AMD(AMD.US)
+7,586%

$Comfort Systems USA(FIX.US)
+4,278%

$Micron Tech(MU.US)
+3,220%

$Arista Networks(ANET.US)
+3,162%

$Lumentum(LITE.US)
+3,011%

$Texas Pacific Land(TPL.US)
+2,835%

$Lam Research(LRCX.US)
+2,534%

$Quanta Services(PWR.US)
+2,396%

$Ciena(CIEN.US)
+2,275%

$Tesla(TSLA.US)
+2,177%

$Axon Enterprise(AXON.US)
+2,067%

$KLA(KLAC.US)
+1,976%

$Broadcom(AVGO.US)
+1,893%

$Monolithic Power(MPWR.US)
+1,635%

First, look at a fact that is most easily overlooked:

👉 Among these 15 companies, most are not "application-layer companies."

Instead, they are:

Semiconductors
Equipment Manufacturing
Network Infrastructure
Power and Engineering Systems

In other words, the answer given by the market over 10 years is:

👉 The ones who consistently create excess returns are the "shovel sellers."

Breaking it down further makes it even clearer.

First Category: Computing Power Core

$NVIDIA(NVDA.US) +19,528%
$AMD(AMD.US) +7,586%
$Micron Tech(MU.US) +3,220%

These companies solve the same thing:

👉 Computing Power

Without computing power, AI would not exist.

Second Category: Semiconductor Equipment and Manufacturing

$Lam Research(LRCX.US) +2,534%
$KLA(KLAC.US) +1,976%

These companies don't make chips directly, but they control:

👉 Whether chips can be manufactured

This is a more upstream segment with higher barriers to entry.

Third Category: Networking and Optical Communications

$Arista Networks(ANET.US) +3,162%
$Ciena(CIEN.US) +2,275%
$Lumentum(LITE.US) +3,011%

These companies solve:

👉 How data flows between data centers

AI is not a single-machine problem; it's a cluster problem.

Fourth Category: Energy and Engineering Infrastructure

$Quanta Services(PWR.US) +2,396%
$Comfort Systems USA(FIX.US) +4,278%
$Texas Pacific Land(TPL.US) +2,835%

Many people overlook this category, but these companies correspond to:

👉 Power + Construction + Physical Systems

Without these, data centers cannot be built.

Finally, a few "System-Level Companies"

$Tesla(TSLA.US) +2,177%
$Broadcom(AVGO.US) +1,893%
$Monolithic Power(MPWR.US) +1,635%
$Axon Enterprise(AXON.US) +2,067%

These companies are more about integration capabilities than a single link.

If you compress this list into one sentence, it's actually very simple:

👉 Over the past 10 years, the best performers weren't the storytellers, but those who control critical resources.

And there's an even more important detail:

These gains are not short-term spikes, but the cumulative result of 10 years.

This means:

What the market truly rewards is "long-term irreplaceability."

So the question is no longer:

Where is the next wave of AI opportunity.

But rather:

In this chain, who holds the "layer that others cannot bypass"?

Are you more inclined towards the next phase's winners continuing to focus on computing power, or will they shift towards the layer of power, networking, and infrastructure?