AI is booming, but also uncertain. Yes, it’s here to stay, but who will lead the industry ten years from now? That’s much harder to predict. New players emerge quickly, and even giants can fall behind in a heartbeat.
I personally invest in AI, but maybe not in the ways you’d expect. In this article, I’ll walk through the different layers of the AI supply chain, give you examples of companies in each, and share my own approach to investing in this fast-moving space.
The AI Supply Chain
AI is not just about the flashy companies building chatbots or autonomous systems. There’s an entire ecosystem powering it, from hardware to software to security. If we simplify it into levels, it might look like this:
Level 1: Producers of machines
Example: ASML (Netherlands) – the only company in the world capable of producing extreme ultraviolet (EUV) lithography systems used to make advanced chips.
- Strengths: Monopoly-like position, huge competitive moat, essential for semiconductor progress.
- Weaknesses: Extremely expensive machines, reliant on a small pool of suppliers, cyclical demand.
Level 2: Producers of chips (foundries)
Example: TSMC (Taiwan) – the world’s largest semiconductor manufacturer, producing chips for Apple, NVIDIA, AMD, and more.
- Strengths: Scale, unmatched expertise in leading-edge chips, global dominance.
- Weaknesses: Heavy geopolitical risk due to Taiwan’s position, high capital intensity.
Level 3: Designers of chips
Example: NVIDIA (USA) – designs GPUs that are critical for AI training and inference.
- Strengths: First-mover advantage in AI chips, strong brand, ecosystem of developers.
- Weaknesses: Valuation is sky-high, competition from AMD, Intel, and custom chips (Google TPU, Amazon Inferentia).
Level 4: AI companies (software & platforms)
Examples: Microsoft, Meta, OpenAI partnerships – companies building the AI applications we interact with directly.
- Strengths: High visibility, potential for huge market adoption, network effects.
- Weaknesses: Intense competition, hype-driven valuations, uncertain regulation.
Level 5: Security against AI abuse
Examples: Cloudflare, Palo Alto Networks – cybersecurity firms protecting against AI-powered attacks.
- Strengths: Growing demand as AI makes cyber threats more sophisticated, established trust.
- Weaknesses: Highly competitive industry, constant need to adapt, sometimes expensive valuations.
How I Invest in AI
Whatever happens in AI, one thing is clear: we will need chips. And that makes the lower levels of the supply chain especially attractive to me.
I’ve invested in ASML, because it has a near-monopoly on advanced chip-making machines. I’m Dutch and It’s a Dutch company, so I hear about it regularly in the news, and it was already on the edge of my circle of competence. That gave me confidence to study it more deeply and eventually invest.
TSMC is also on my radar, but I’ve been cautious given Taiwan’s political risks and its stretched valuation.
At Level 3 (chip designers), things get tricky. The landscape changes fast, and today’s leader can quickly become tomorrow’s loser. I’m wary of betting too heavily on one horse.
The only Level 4 (AI software) company I’ve felt comfortable investing in is Microsoft. Its business is diversified, cloud, enterprise, Office, gaming, and AI is just one piece of the puzzle. That means AI growth feels like a bonus, not the whole story.
Finally, I like Level 5 (security) companies, because as AI grows more powerful, so do the risks. Criminals will use it, and demand for cybersecurity will only increase. This feels like a quieter but crucial growth story.
What I avoid:
- Startups with no proven business model.
- Businesses I can’t explain in simple terms, if I can’t describe it to a ten-year-old, I probably shouldn’t invest.
Conclusion
AI is one of the most exciting investment opportunities of our time, but also one of the most uncertain. Instead of trying to guess which company will build the “next ChatGPT,” I prefer looking at the picks-and-shovels businesses that enable AI to exist in the first place. Chips will be needed no matter what. Security will be needed no matter what.
For me, that means companies like ASML, TSMC, and Microsoft form the safer core of my AI exposure. They might not double overnight, but they have moats and staying power.
If you’re thinking about investing in AI, start by asking yourself:
- Which parts of the supply chain do I actually understand?
- Am I chasing hype, or am I buying something that will still matter in 10 years?
AI may be about the future, but your investment decisions should be grounded in the present.
Learn more about your Circle of Competence here, or how ro research a company before buying.