From AI infrastructure to semiconductor supply chains – these are the companies I’m considering adding to my portfolio.

The biggest investment opportunities over the next decade may not come from companies building AI models, but from those supplying the infrastructure that makes artificial intelligence possible.

After building core positions in NVIDIA, AMD, Broadcom, Alphabet, Meta, Micron and Rocket Lab, I’m now looking at the next layer of the AI ecosystem.

These are the eight companies currently on my watchlist.

---

1. Credo Technology (CRDO) ⭐⭐⭐⭐⭐

Why I like it

If NVIDIA is building the world’s most powerful GPUs, Credo is helping those GPUs communicate at incredible speeds.

As AI clusters continue expanding, demand for high-speed connectivity is rising rapidly.

The company recently reported another quarter of record revenue growth, driven by accelerating demand for AI networking products.  

Pros

✔️AI networking leader

✔ Massive AI data centre exposure

✔ Strong revenue growth

✔ High gross margins

✔ Potential multi-year winner

Risks

• Premium valuation

• Customer concentration

• High volatility

My rating: 9.8/10

---

2. Fabrinet (FN) ⭐⭐⭐⭐⭐

One of Wall Street’s best-kept secrets.

Rather than designing chips, Fabrinet manufactures the advanced optical components used by many of the biggest networking companies.

As AI infrastructure spending increases, Fabrinet benefits regardless of which networking company wins.

Pros

✔ AI optical manufacturing

✔ Strong profitability

✔ Excellent execution

✔ Less competition than many semiconductor firms

Risks

• Slower growth if AI spending cools

• Hardware manufacturing remains cyclical

My rating: 9.5/10

---

3. Sterling Infrastructure (STRL) ⭐⭐⭐⭐☆

Everyone talks about GPUs.

Almost nobody talks about who builds the data centres.

Sterling benefits from:

• AI data centre construction
• Energy infrastructure
• Industrial expansion

Sometimes the biggest winners are the companies building the roads rather than driving on them.

Pros

✔ Infrastructure demand

✔ Strong earnings momentum

✔ Beneficiary of US investment spending

Risks

• Construction cycle

• Government spending risk

My rating: 9.3/10

---

4. Ichor Holdings (ICHR)

Semiconductor equipment doesn’t work without suppliers.

Ichor provides fluid delivery systems used in advanced chip manufacturing equipment.

If TSMC, Samsung and Intel continue expanding capacity, suppliers like Ichor could also benefit.

Pros

✔ Semiconductor equipment exposure

✔ AI manufacturing cycle

✔ Attractive valuation

Risks

• Highly cyclical

• Dependent on semiconductor capital expenditure

My rating: 9.0/10

---

5. StoneX (SNEX)

Unlike the others, StoneX isn’t an AI company.

It’s a diversified financial services business with growing earnings and exposure to global commodities and trading.

A good diversification candidate.

Pros

✔ Consistent earnings

✔ Financial services growth

✔ Lower volatility

Risks

• Limited AI upside

• Financial market sensitivity

My rating: 8.8/10

---

6. Lumentum (LITE)

Optical networking will become increasingly important as AI clusters scale.

Lumentum already supplies advanced optical technology used across cloud infrastructure.

Pros

✔ AI optical demand

✔ Data centre exposure

✔ Recovery potential

Risks

• Fierce competition

• Margin pressure

My rating: 8.8/10

---

7. SanDisk (SNDK)

AI doesn’t only require GPUs.

It also needs enormous amounts of fast storage.

Demand for enterprise SSDs could remain strong throughout this decade.

Pros

✔ AI storage demand

✔ Enterprise SSD exposure

✔ Well-known technology brand

Risks

• Memory pricing cycles

• Highly competitive market

My rating: 8.7/10

---

8. Dycom (DY)

AI requires more fibre.

Dycom builds telecommunications infrastructure that supports expanding network capacity.

Pros

✔ Fibre expansion

✔ Stable infrastructure demand

✔ Long-term connectivity trend

Risks

• Lower growth than AI chip companies

• Capital spending cycles

My rating: 8.6/10

---

My Personal Ranking

🥇 Credo Technology (CRDO)

🥈 Fabrinet (FN)

🥉 Sterling Infrastructure (STRL)

4️⃣ Ichor Holdings (ICHR)

5️⃣ StoneX (SNEX)

6️⃣ Lumentum (LITE)

7️⃣ SanDisk (SNDK)

8️⃣ Dycom (DY)

---

Final Thoughts

The next decade of AI won’t be powered by a single company.

It will require chips, networking, optical components, storage, electricity, construction and fibre infrastructure.

That’s why I’m increasingly looking beyond the obvious names. My core holdings remain in large AI leaders, but these eight companies could offer additional exposure to the industries enabling the next generation of AI infrastructure.

This article reflects my personal research and investment watchlist, not financial advice.