AI is a chain of industries, not one stock
When people talk about investing in AI they often mean one or two famous chip companies. In reality the money flows through a chain. Someone designs the chips, someone manufactures them, cloud companies buy them by the hundred thousand, other firms build and power the data centres, and software companies turn it all into products people pay for.
Each link earns money in a different way and carries different risks. Chip sales can swing sharply from year to year. Cloud platforms earn steadier income but spend enormous sums up front. Software companies depend on customers actually paying for AI features.
Check what you already own
The largest AI-related companies are also among the largest companies in the world. If you hold a global or US tracker fund, a meaningful share of it is already in them.
Before adding an AI investment, look at the top ten holdings of the funds you own. You may be more exposed to AI than you think, and adding more makes your investments less spread out, not more.
Decide how concentrated you want to be
A portfolio of three chip companies and a portfolio of twenty companies across the whole chain are both AI investments, but they behave very differently. The first moves almost entirely with chip demand. The second is steadier and less dependent on any one result.
Setting a maximum for any single company, for example 8% or 10%, is a simple way to stop one name dominating.
Think in years, not weeks
AI shares have moved a long way in both directions over short periods. Anyone investing in a single theme needs to be able to sit through large falls without being forced to sell.
Investing a fixed amount regularly, rather than everything at once, is a common way to avoid putting all your money in at a high point.
Who does what in AI
Chip designers
Design the processors that AI systems are trained and run on.
For example: Nvidia, AMD, Broadcom
Chip manufacturing
Make the chips, and the machines used to make them.
For example: TSMC, ASML, Applied Materials
Cloud platforms
Run huge data centres, rent out computing power and build their own AI models.
For example: Microsoft, Alphabet (Google), Amazon
Data-centre equipment and power
Supply the networking, cooling and electrical equipment data centres need.
For example: Arista Networks, Vertiv, Eaton
Software and applications
Sell AI features and tools to businesses and consumers.
For example: Salesforce, ServiceNow, Palantir
Companies are listed to show what each category means. They are not recommendations, and the list is not complete.
Three ways to invest
There is no single right route. Many people combine them, and plenty decide a broad global fund is all they need.
- 1
Buy individual shares
Pick one or more AI companies and buy their shares through a broker or investing app.
- Good for
- Simple to understand. You own exactly what you chose.
- Watch out for
- If you only hold a few companies, one bad result can do real damage. Picking winners is hard, even for professionals.
- 2
Buy a ready-made fund
A fund (often an ETF) holds dozens of companies in one purchase. Examples include the Global X Artificial Intelligence & Technology ETF (AIQ) in the US and the WisdomTree Artificial Intelligence UCITS ETF (WTAI) in the UK and Europe.
- Good for
- Instant spread across many companies, with very little effort.
- Watch out for
- You get what the fund provider chose, including companies you may not want. Themed funds usually charge more than broad trackers, so check the yearly fee and the top ten holdings. Which funds you can buy depends on where you live.
- 3
Build your own portfolio
Choose your own mix of companies and how much of each, then buy them through your broker. This is what Arithmos helps with: describe the mix in a sentence and it builds and tests one for you to consider.
- Good for
- You decide exactly what is in and what is out, and you can see the reason for every company.
- Watch out for
- More to look after than a fund. Buying many separate shares can cost more in dealing fees, and a tested result is a simulation, not a promise.
The risks
- Concentration. A handful of companies dominate the theme, so problems at one can drag down the whole group.
- High expectations. Many AI share prices assume years of strong growth. If growth is merely good rather than exceptional, prices can still fall.
- The spending cycle. Chip and equipment sales depend on large technology companies continuing to spend heavily on data centres. That spending can slow.
- Private companies. Not every important AI company is listed on a stock market, and private companies cannot be bought through an ordinary broker.
- Regulation and competition. Rules on AI, export controls on chips and new competitors can all change the picture quickly.
Try it yourself
The quickest way to understand a portfolio is to look at one. Here is how to turn this guide into something you can inspect.
- 1
Start from the idea
The button below opens the builder with this guide's idea already typed in. Change any part of it: the number of companies, the countries, the limit per company.
- 2
Read the reason for every company
You get a list of companies, how much of each, and a plain-English reason for each pick. If one looks wrong to you, you can ask for it to be changed.
- 3
Look at the bad years, not just the good ones
The portfolio is replayed against real past prices. Check the biggest fall along the way, and ask yourself whether you could have sat through it.
- 4
Decide for yourself
If you want to act on it, you buy the shares yourself through your own broker. Arithmos never holds your money or places trades.
“Companies across the AI supply chain: chip designers, chip factories, cloud platforms, data-centre equipment and AI software. US-listed, 20 companies, no single one over 8%.”
Browsing portfolios other people have published is free. See pricing for what building your own includes. A tested result is a simulation using past prices. It is not a forecast and not a recommendation.
Common questions
What is the easiest way to invest in AI?
The lowest-effort route is a fund that holds many AI-related companies, bought through a broker or investing app. A broad global or US tracker fund also gives you meaningful exposure, because the largest AI companies are among its biggest holdings.
Can I invest in AI with a small amount of money?
Yes. Many brokers let you buy fractions of a share or small amounts of a fund, so you do not need the full price of an expensive share.
Is it too late to invest in AI?
Nobody can know. Prices already reflect high expectations, which means future returns depend on companies doing even better than investors currently assume. That uncertainty is a reason to spread your money and to avoid investing more than you could bear to see fall sharply.
Which AI stocks should I buy?
Arithmos does not give personal recommendations. This guide explains how the industry is structured so you can judge for yourself, or with a qualified adviser, what suits your situation.
