An analysis of the S&P 500's ten-year dynamics reveals a striking pattern: nine out of the ten growth leaders are tied in one way or another to artificial intelligence infrastructure. The undisputed king of this race is Nvidia.

Nvidia shares have delivered investors nearly 13,589% returns over the past ten years—more than double the pace of its closest rival, AMD, with its impressive ~6,000%. The top ten also includes chipmakers and networking equipment manufacturers, and even one HVAC contractor.

Full list of winners

  • Nvidia (NVDA) — +13,817%
  • AMD (AMD) — +6,099%
  • Micron (MU) — +5,486%
  • Comfort Systems (FIX) — +5,157%
  • Arista Networks (ANET) — +3,762%
  • Lam Research (LRCX) — +3,099%
  • Tesla (TSLA) — +2,545%
  • Lumentum (LITE) — +2,440%
  • KLA Corp (KLAC) — +2,401%
  • Seagate (STX) — +2,346%

Nvidia, AMD, Micron, Lam Research, and KLA Corp are either chips themselves or equipment for their production. These components are the foundation of servers for data centers, which underpin all AI development. Arista Networks supplies network switches for the same data centers, Lumentum produces optical components for data transmission between racks, and Seagate provides storage drives for training neural networks.

Comfort Systems has its own unique niche. This engineering and electrical systems contractor has accumulated nearly $12 billion in orders: companies are mass-building and cooling new data centers, and Comfort Systems is profiting from this boom without selling a single chip. Tesla is the exception—its growth is driven primarily by demand for electric vehicles, but Elon Musk's projects in autopilot and robotics still tie the company to AI.

The last two years brought the most profit

The key point: growth was not uniform. The main surge came in the final stretch of the decade. Since the start of the AI boom in November 2022, Nvidia's market capitalization has jumped from $418 billion to over $4.5 trillion. Other companies in the top ten made similar leaps—hyperscalers began spending enormous sums on AI.

Future dynamics directly depend on whether they maintain current construction pace. By my estimates, about 60% of data centers planned by 2027 have not even started construction yet. This lag could keep investor interest in such stocks high well into the second half of the decade.

My take: we are witnessing not just cyclical growth but a structural shift. AI infrastructure is a new "golden age" for the semiconductor industry, and those betting on "bricks and cooling" for data centers may benefit just as much as chipmakers.