3 types of companies that could profit significantly from AI agents. And 2 where I'd be more cautious.
Meta launched Muse only on September 8, and in the first 12 days it got around 2.8 million downloads. More important, though, is what it can do. It no longer just answers, but can send emails, book travel, or make a purchase.
And that could significantly change the internet economy.
Until now, a person searched for a product, compared, and then bought. An AI agent can do all three steps on their behalf.
1. Shopify $SHOP
For me, one of the clearest cases. The biggest problem of a small e-shop has always been getting in front of the customer. But an AI agent can search a huge number of products and find the one that best meets the brief on its own.
Muse has already gained access to the entire catalog of stores on this platform, and Meta is also linking purchases through Shop Pay. The company doesn't even have to own the AI agent itself. What matters is that it can provide the technology and the commercial infrastructure through which the agent completes the purchase.
An agent can bypass many of today's intermediaries, but someone still has to execute the payment. And with AI, there's even greater emphasis on who verifies identity, approves a specific amount, and protects the customer from fraud.
Visa is already working with OpenAI on payments made by AI agents, and Mastercard has already carried out real transactions where the agent completed the purchase itself within set permissions. So if AI increases the number of automated online transactions, Visa and Mastercard can continue to collect their small "toll" from a huge volume of payments.
3. Micron $MU
Here the link to AI agents is a bit less direct.
An AI agent won't do just one question → one answer. It must plan, work with long context, use additional tools, store information, and perform multiple steps. That means growing demands not only on computing power but also on DRAM, HBM, and storage.
Micron directly states that AI agents are already increasing server demand and that this type of AI is changing data centers, meaning servers for managing agents are being added as well as storage for their growing context. That's why memory seems to me one of the most interesting indirect ways to benefit from this trend.
And who would I be more cautious about?
Here it should be stressed that these are not companies that are automatically threatened. Rather, these are business models where AI agents create new risk. These companies can adapt, and some of them are already doing so.
4. Booking $BKNG and Expedia $EXPE
Their value is largely in gathering offers in one place and allowing comparison. But the same hotel room can often be found on multiple sites. An AI agent can go through all the options itself and book where the conditions are best. That potentially weakens the value of the intermediary itself.
On the other hand, we already see how these firms are trying to adapt.
Expedia is already trying to use this trend to its advantage and works directly with Muse. Instead of remaining just a platform that an agent can bypass, it wants to be one of the sources of offers from which the agent will choose. That's why I wouldn't write off Booking or Expedia yet. The question is more: who will own the customer relationship in the future and who will keep a bigger share of the margin?
5. Alphabet $GOOGL
Here I see perhaps an even more interesting risk. Classic AI has taken a slice of informational search. But an AI agent can hit commercial search, which is much more profitable. Instead of a person searching for a product, going through results, and clicking on ads, they can simply say:
"Find me the best option under $150 and buy it."
Part of today's path to purchase simply disappears. But Google shouldn't be written off here either. It has a huge Shopping Graph with more than 60 billion product listings and is already building its own shopping features for AI agents. So the question is: Can Google monetize AI agents as effectively as it does classic search today?
So for me the main point:
AI agents can reduce the value of companies whose main advantage is getting the customer to the product. Conversely, they can strengthen companies that own the infrastructure needed for the purchase to happen at all.
Which company do you think will gain the most from the rise of AI agents? 👇
