How to Invest in AI
Last updated June 2026
Short answer
To invest in AI, first understand that the theme has layers (chips and semiconductors, infrastructure and data centers, models and platforms, and applications and software), and then choose a vehicle. There are three main ways: buy individual AI stocks across the stack for maximum control, buy an AI or broad-tech ETF for diversified exposure in one ticker, or build a thematic basket you manage yourself. Each has honest trade-offs, and the real risks (hype, stretched valuations, and concentration in a few large names) are worth taking seriously. Size AI as one slice of a diversified portfolio, not the whole thing. Walnut can help you build and manage an AI basket you approve, but Walnut is not an investment adviser.
“How do I invest in AI?” sounds like one question, but it is really two: which part of the AI theme do you want to own, and what vehicle do you use to own it? AI is not a single company or product; it is a stack, from the chips at the bottom to the apps at the top, and value shows up at different layers at different times. This guide lays out those layers, walks through the three practical ways to get exposure (individual stocks, ETFs, and a thematic basket), is honest about the trade-offs and the real risks, and covers how to size a position. It is a description of the options, not a recommendation to buy anything.
The layers of the AI theme
Before picking a stock or a fund, it helps to know where in the stack you are buying. “AI” spans four rough layers, and a company’s layer shapes both its opportunity and its risk. Spending tends to flow into the lower layers (chips and infrastructure) first, before the applications on top turn it into revenue.
Chips and semiconductors
The hardware layer: the processors, accelerators, and memory that AI models are trained and run on, plus the equipment and design tools that make those chips. This is the most direct “picks and shovels” exposure to AI demand.
Chip designers, foundries, memory makers, and semiconductor-equipment suppliers.
Infrastructure and data centers
The buildout layer: the data centers, cloud platforms, networking, power, and cooling that AI workloads run inside. AI is enormously compute-hungry, so a lot of spending flows here before any application makes money.
Cloud providers, data-center operators, networking, and the power and cooling that feed them.
Models and platforms
The intelligence layer: the companies building the large models and the platforms that serve them. Some are public, many of the pure-play labs are private, so public exposure here often runs through their larger backers.
Large public platforms building or hosting foundation models; many pure-play labs are still private.
Applications and software
The product layer: software companies putting AI into tools people actually use, from productivity and search to security and industry-specific apps. This is where AI has to translate into revenue rather than spending.
Software and internet companies embedding AI features into products and workflows.
You do not have to own every layer, but knowing which one a holding sits in is how you avoid accidentally concentrating your whole AI position in a single part of the stack. For a deeper look at the buildout layer, see the AI infrastructure theme.
The three ways to invest in AI
Once you know which layers interest you, you pick a vehicle. There are three practical routes, and they trade off control against effort and diversification. None is the single right answer; the best one depends on how much you want to research and manage.
Individual AI stocks
Buying shares in specific companies across the stack: a chipmaker, a cloud provider, a software name. You choose exactly where in the layers you want exposure and how much of each.
- Best for: Investors who want precise control over which companies and which layers they own, and are willing to research each name.
- The trade-off: Single-stock risk is real: one company can fall hard on an earnings miss or a lost customer, and picking winners across a fast-moving theme is genuinely difficult.
AI and tech ETFs
Buying a fund that holds a basket of AI-related or broad-technology companies in one ticker. Some track a narrow AI theme, others are broad tech or the whole market with heavy tech weight.
- Best for: Investors who want diversified exposure to the theme without picking individual names, in a single, low-effort holding.
- The trade-off: You own the fund’s definition of “AI,” which can be loose or overlap heavily with what you already hold, and you pay an expense ratio for the convenience.
A thematic basket
Building your own small, deliberate group of AI holdings across the layers, with a written thesis and target weights you manage over time. It sits between a single ETF and a pile of unrelated single stocks.
- Best for: Investors who want a considered, diversified-within-the-theme position they understand and control, rather than a black-box fund or scattered bets.
- The trade-off: It takes more thought than buying one ETF, and a self-built theme concentrated in one sector still carries the concentration risk of that sector.
Many investors combine these: a broad-tech or AI ETF as a base, a few individual names they have conviction in, and a written thesis that keeps it all deliberate. For how the fund route works in detail, see the best AI ETFs guide, and for the basket approach see thematic investing.
At a glance
| AI layer / way to invest | Best for | Trade-off |
|---|---|---|
| Individual AI stocks | Investors who want precise control over which companies and which layers they own, and are willing to research each name | Single-stock risk is real: one company can fall hard on an earnings miss or a lost customer, and picking winners across a fast-moving theme is genuinely difficult. |
| AI and tech ETFs | Investors who want diversified exposure to the theme without picking individual names, in a single, low-effort holding | You own the fund’s definition of “AI,” which can be loose or overlap heavily with what you already hold, and you pay an expense ratio for the convenience. |
| A thematic basket | Investors who want a considered, diversified-within-the-theme position they understand and control, rather than a black-box fund or scattered bets | It takes more thought than buying one ETF, and a self-built theme concentrated in one sector still carries the concentration risk of that sector. |
The risks: hype, valuations, and concentration
AI is a real and large trend, but a real trend and a good investment at today’s prices are not the same thing. Three risks are worth taking seriously before you buy anything:
- Hype. Expectations for AI are high, and enthusiasm can push prices ahead of what the businesses have actually delivered. When a story is this popular, a lot of good news may already be priced in, which leaves less upside and more room for disappointment.
- Stretched valuations. Some AI-linked stocks trade at valuations that assume years of strong growth. That can work out, but it also means a mild disappointment can trigger a sharp drop, because the price was leaning on optimistic assumptions. It is worth understanding what growth a valuation implies before buying.
- Concentration. The theme is dominated by a handful of very large companies, and those same names are often the biggest holdings in broad index funds. So “adding AI” can quietly mean doubling down on stocks you already own heavily, turning what feels like diversification into concentration.
None of this means avoid AI. It means going in with clear eyes: check your existing exposure, spread across the layers rather than one name, and do not assume the trend continues in a straight line.
How to size an AI position
Sizing is where most of the risk is actually managed, and it is personal to your goals, timeline, and risk tolerance. This is a description of common approaches, not advice:
- Treat it as a slice, not the whole. A single theme usually works as one part of a diversified portfolio rather than the centerpiece. Size it so a bad year for AI would be uncomfortable but not damaging to your overall plan.
- Count what you already own. Before adding, check how much AI exposure your index funds already carry through their largest holdings, so you are not unknowingly concentrating.
- Spread across the layers. Within your AI slice, holding chips, infrastructure, platforms, and applications spreads the bet across where value can accrue, instead of resting it on one company or one part of the stack.
- Decide the thesis and the exit up front. Write down why you are buying and what would change your mind. A thematic basket with target weights makes this concrete and easier to manage over time.
Where Walnut fits
To be upfront, since this is our site: Walnut is one option among many, not the only way to invest in AI, and not a number-one pick. Walnut is an AI investing assistant you chat with on the broker you already own. It connects your existing brokerage through SnapTrade (read-only by default), and lets you research the AI theme and build a thematic basket across the layers, with target weights and a written thesis you control.
The distinctive part is that the conversation is grounded in your real holdings, and each position is framed against the S&P 500 so you can see how a name is doing relative to just owning the index. Walnut builds the basket, but you approve it, and every trade needs your confirmation at your own broker. It is not a fund, not a robo-adviser, and not a stock picker: it helps you research and structure a position you understand. Walnut is informational and is not an investment adviser; the thesis and the decision are yours.
The bottom line
Investing in AI comes down to two choices: which layers of the stack you want (chips, infrastructure, models, or applications) and which vehicle you use (individual stocks, an ETF, or a thematic basket). Each vehicle trades control against effort and diversification, and none is universally best. Take the risks seriously, hype, stretched valuations, and concentration in a few large names, and size AI as one considered slice of a diversified portfolio rather than an all-in bet. Walnut can help you build and manage an AI basket you approve, but it is one option, and it is not an investment adviser.
For the vehicles in more depth, see the best AI ETFs and how thematic investing works, or explore the AI infrastructure theme.
Try Walnut on top of your broker
Walnut connects any major US broker in a few clicks, then lets you research the AI theme and build a basket you approve, with each holding framed against the S&P 500. Read-only by default; you approve every trade.
FAQ
How do I invest in AI?
There are three main ways: buy individual AI stocks across the stack (chips, infrastructure, models, applications), buy an AI or broad-tech ETF for diversified exposure in one ticker, or build a thematic basket you manage yourself. There is no single right answer; it depends on how much control and effort you want. Whatever route you pick, size the position sensibly and remember this is a description of the options, not advice. Walnut is not an investment adviser.
What are the layers of the AI theme?
It helps to think of AI as a stack. Chips and semiconductors are the hardware that trains and runs models. Infrastructure and data centers are the compute, cloud, power, and cooling that everything runs inside. Models and platforms are the companies building the large models. Applications and software are the products that put AI in front of users. Different companies sit at different layers, and spending shows up in the lower layers first.
Is it better to buy AI stocks or an AI ETF?
Neither is universally better. Individual stocks give you precise control over which companies and layers you own, at the cost of single-stock risk and more research. An AI ETF gives you diversified exposure in one ticker, at the cost of owning the fund’s definition of AI and paying an expense ratio. Many investors use both, or build a thematic basket that sits in between. Match the vehicle to how much control and effort you want.
What are the risks of investing in AI?
The main ones are hype, valuation, and concentration. Expectations for AI are high, so prices can already reflect a lot of optimism, and stretched valuations can fall sharply if growth disappoints. The theme is also concentrated in a handful of very large companies, so “AI exposure” can quietly mean a big bet on a few names you may already own through index funds. Diversifying across layers and sizing the position modestly are the usual ways to manage this.
How much of my portfolio should be in AI?
There is no universal number, and this is not advice. A common approach is to treat a single theme as a slice of a diversified portfolio rather than the whole thing, and to check how much AI exposure you already have through broad index funds before adding more. Size it so that a bad year for the theme would be uncomfortable but not damaging to your overall plan. Your own goals, timeline, and risk tolerance decide the figure.
Are AI stocks overvalued in 2026?
It depends on the company and who you ask, and no one can tell you with certainty. Some AI-linked stocks trade at high valuations that assume years of strong growth, which leaves less room for error if results disappoint. Others are more modestly priced. The honest answer is that a lot of optimism is already reflected in many prices, so it is worth understanding what growth a valuation implies before buying, rather than assuming the trend continues.
Can I invest in AI without buying individual stocks?
Yes. AI and broad-technology ETFs let you get exposure to the theme in a single ticker without picking individual companies, and broad market index funds already carry meaningful AI weight through their largest holdings. This is the lower-effort route and spreads single-stock risk, though you take on the fund’s definition of AI and its expense ratio. See our guide to the best AI ETFs for how these funds differ.
What is a thematic AI basket?
A thematic basket is a small, deliberate group of AI holdings you choose across the layers, tied to a written thesis and target weights you manage over time. It sits between a single ETF and a scattered pile of single stocks: more control and understanding than a black-box fund, more structure than random bets. Walnut can help you build one from your existing brokerage, with each holding framed against the S&P 500, but the thesis and the decision are yours.
How do I diversify within AI?
Spread across the layers instead of loading up on one. Owning only chipmakers, for example, is a concentrated bet on one part of the stack; adding infrastructure, platform, and application exposure spreads the theme across where value can accrue. Also check overlap: several AI names may already be large positions in your index funds, so “diversifying” into them can actually concentrate you further.
Does Walnut pick AI stocks for me?
No. Walnut is an AI investing assistant you chat with on the broker you already own. It can help you research the AI theme, build a thematic basket, and frame each holding against the S&P 500, but it connects read-only by default and you approve every trade. It describes options and trade-offs rather than issuing recommendations. Walnut is informational and is not an investment adviser; the decisions are yours.
How do I actually place AI investments?
You buy through a brokerage account, the same way you would buy any stock or ETF. You can do it directly in your broker’s app, or use a tool that sits on top of your existing broker. Walnut connects your brokerage through SnapTrade (read-only by default), lets you build an AI basket you approve, and places the trades at your own broker once you confirm. Whichever route you use, the account and the money stay yours.
Is investing in AI a good idea?
That depends on your goals, timeline, and risk tolerance, and no one can promise it will work out. AI is a genuine, large trend, but a strong trend and a good investment at today’s prices are not the same thing, because a lot of optimism can already be reflected in valuations. Treating it as one considered slice of a diversified portfolio, rather than a concentrated all-in bet, is a common way to participate without overexposing yourself.
Walnut is informational and is not an investment adviser. Markets, valuations, funds, and availability change; verify current details before deciding. Nothing on this page is a recommendation to buy, sell, or hold any security or to use any particular product.