持有 GOOGL 可以接受，但并非有吸引力的买入标的；AI 颠覆风险被强劲的基本面和合理的估值所抵消。

Alphabet's stock was one of the biggest winners in 2025, but it faced slightly greater difficulties in 2026. The stock fell 14 % from its peak.

The question is, will this be a company that will be disrupted by some of the new technologies coming from agents and artificial intelligence? John, one of the things I wanted to start with regarding Alphabet.

It is a very interesting company to follow because it is a huge technological force . I think they have, perhaps more than anyone else, cash flow. They started borrowing at this stage.

But if you look at everything they have in their ecosystem, they basically own everything starting with the models,

They were really at the forefront of developing many AI models, and they developed TPU modules, but it seems they are starting to fall behind. So, it seems like we're looking for an early warning sign of a possible disturbance at Alphabet? This is where we should kind of start.

Yes, with regard to the disruption, we can say at the very least that it has not done enough to lead the way, and there is an argument that can be made that no one can do enough to lead the sector sustainably because this field is rapidly evolving and highly competitive.

So, in the best-case scenario , it's really hard to see Google as a sustainable leader when it comes to this.

It is certainly part of the conversation, but what makes it difficult is that $200 billion that is spent annually as capital expenditure on building artificial intelligence. What are the expected economic returns, and is it sustainable at all? It could be argued that it is not .

Yes, those capital expenditures are insane, and that's part of Alphabet's dynamic that I think is hard for us to grasp from the outside, which is that you have a huge cloud business that's growing like crazy. I mean, I think it's above 50% at this stage.

But you also have internal needs. So the two big customers to consider here with Alphabet are actually Anthropic, which has trained all of its models so far on TPU units. Essentially, in the Google cloud.

So this is a customer who pays money to Alphabet. Alphabet also owns a stake in it . But they pay them money.

But they are working on building this infrastructure that internal developers want to use to develop the next generation of Gemini. So, from everything we've heard, there is real tension in this regard.

One reason they may not be at the forefront of modeling is that they are giving these computing capabilities to Anthropic, which is almost like funding their competitors.

Yes, and from an outside perspective, you might wonder: why don't they stop doing that? It's like a trap. What is the alternative? It is clear that everything is moving towards artificial intelligence, so you want to be a company that builds this field and seeks to lead the way.

But doing so means you are moving away somewhat from what your job and primary source of income have been for many years . But I do n't really know what the alternative is, because I believe that all technology will be built on the basis of artificial intelligence, and to some extent, for better or for worse, that is the path Alphabet has to follow.

But I think one thing is clear, which is that Alphabet has not been a leader when it comes to these agents. This is very interesting because they have some really strong potential points of contact.

In particular, the Android system. Since more than half of the world's smartphones run on Android, you can imagine the possibility of them having smart agent capabilities, whether through voice, through the device in your pocket, through messaging , or in some other way. Therefore, they can build this type of product.

They also have the cloud, which is something Meta built, but Alphabet's cloud is bigger , and arguably works better, because that's what they've been building to run the search engine for a long time.

They also have all sorts of other contact points such as Gmail and Google Calendar.

I mean, I initially linked "Muse" to some Google products , and this is how I use it . But is there a possibility that they simply missed this agent business model? Or perhaps they are stuck in a world where they think about things from a vertical integration perspective , wondering how to get their agents to work with their calendar and email, but not thinking about how to work with a platform like Shopify.

It is a very strong connection to the past. If you look at Google or Alphabet, you'll find that two-thirds of their revenue base comes from advertising, digital advertising. This is something that agent-based commerce and interactive artificial intelligence can significantly disrupt.

Sure, I mean you have YouTube with its unskippable ads , so at least that's there. But when it comes to the open internet and the digital ecosystem out there, the agents don't click on ads.

So, I was having my Muse agent look for things. I think if you're thinking about dealer-based commerce , we're at the "iPhone 1" stage of the iPhone journey. Things will only get better from here.

So, in the future , I truly believe that people will delegate shopping tasks to their agents. There are some questions that need to be resolved, but I believe this will happen, and it will significantly disrupt one of the greatest businesses of all time: Google's advertising business.

Do you think that having Android, and having that operating system, as we've seen them expand Android to include things like robots, would be a good place for that to be an operating system platform ?

Because it seems that if there is a flaw in Muse, it is that it is just a phone app at the moment. They are trying to get it to the point where it has devices like Tamagotchi; They develop open-source software that you can actually put into other pieces of hardware, but they would like to be in a position where you interact with a Muse device, not an iPhone or Android phone.

If we live in a world where an iPhone or Android phone is the primary point of interaction , it seems they can enter this field, or am I overthinking and giving them more credit than they deserve?

Well, I'll say that the most useful way I've interacted with Muse so far has been via WhatsApp, by simply sending him a voice message. One day, I was taking a walk and I let out everything that was going through my mind.

Here is everything I have in mind that I need to do during the next week. I don't know when I will do it, but I know when it needs to be done. It was an audio clip that was about 3 minutes long , and Muse was able to turn it into a to- do list with the days on which things should be done.

So, this is "Muse" actually living on my device in a meaningful and accessible way.

But I will say this: Yes, having an operating system that is not restricted by a closed ecosystem can be very beneficial, because I think one of the great things we will see in agent-based commerce in the future is that it will take some of the power out of the big markets and give it to the small companies operating in them.

This will make it more discoverable and more connected to the consumer. Therefore, having an open system and an open ecosystem can be very powerful.

Let's move on to research, because I really think it's a very interesting field . Research has always been the driving force behind Alphabet's profitability since the company was actually founded.

I mean, in the early 21st century they really figured out how to profit from this, and it has continued to build up for decades now. We haven't seen any disruption in the search numbers yet, but I think we're starting to see signs of movement from those companies that were better at generating profits than we were in the past, especially Alphabet.

For example, you might conduct research and generate profits, let's say from an insurance advertisement. You say, "Hey Google, I want to do something." " I want to spend some money."

Then Google matches you with a customer, and they want to match you with the best customer, and also the best customer who pays the highest price. That has always been the secret to their success.

It wasn't just about the higher price, but also about the click-through rate and what that meant for making profits.

If we were working in a world of smart agents and I said, for example, " Hey Muse, or any other agent, I'm looking for new home insurance for next year." What are my available options?

Perhaps you can provide me with price quotes immediately. There may be no advertisements in that context. Do you believe there is a real threat to research activity and ways of profiting from it?

Or will they eventually find a way to make a profit, as they did with features like "AI mode " in search?

If you go into the search engine now, you will find this AI-powered option. We may live in a world where you use your browser as usual, but it looks a little different than it did five years ago.

To clarify, Alphabet is one of the most powerful companies ever, and enjoys one of the broadest competitive advantages in history. So, on one hand, Alphabet is definitely very powerful .

We are looking at the other side of this equation . What threatens this business activity? While thinking about these things is fun and interesting, I believe Alphabet should be given high credit for its ability to adapt and profit in a changing technology landscape.

However, let's be clear that the technological landscape is indeed changing, and we are seeing things emerging in research right now that are further complicating this activity.

One of those things is AI-powered research, like what I just talked about regarding " AI summaries ". A Pew Research Center study earlier this year found that 26% of people end their research prematurely because of these summaries.

Not long ago, this percentage was 16%. Therefore, there is a large gap that is now widening, as AI summaries are bringing the search process to an early end. If you finish your search too early, it means ending the opportunity for Google to generate revenue.

Another study conducted by SparkToro earlier this year also found that 68% of searches ended without any clicks. That figure was only 60% of searches two years ago. In the big picture, this is indeed a big leap and coincides with artificial intelligence increasingly entering the scene.

People now interact with chatbots to find the information they want instead of using a traditional search engine . This significantly changes Google's profitability opportunities, and this trend is likely to accelerate from now on.

Do you think they will be able to understand advertising in a world of smart agents?

So, is this a possible disorder and is there a possibility of replacement, but perhaps that replacement is not of the same quality? This applies to both companies, doesn't it? Perhaps the future of smart agents is not as profitable as the current model.

I've turned this over and over in my head , and it's absolutely true. I believe that the world of intelligent agents disrupts the current paradigm, and this alternative paradigm is not as valuable as what we leave behind.

This is what I come to time and time again, because I really don't see how they will be able to make a profit like they did in the past.

Well, I want to connect this product-related aspect , because Alphabet is one of my biggest investments, and the reason I keep pushing is whether it should remain one of my biggest holdings.

I'll get to the evaluation here before we finish, but the most important part I'm thinking about is whether they can really innovate new products? And I think the "Muse" moment was really educational, because it showed that maybe it's because Mark Zuckerberg is in charge, or maybe it's because everything comes from him, right?

He hired the people who ended up creating "Muse" and they were able to rethink how things worked in the field of artificial intelligence for their business. But does Alphabet possess this power?

Do they have the ability to create a new product or even quickly follow up on something?

I'm just thinking, I'd much prefer having an agent from Alphabet that I could link to my email and calendar to manage family affairs. This is what I find really useful at the moment.

I really don't want to hand that over to Meta because then I would be faced with two of the biggest technology companies looking into the details of my life.

Ideally, all of that should be within a single ecosystem. But they simply did not do that at this stage. I know you can do these things if you want to delve into the finer details.

But turning technology into products, I think they are still an amazing technology company, but turning that technology into a product seems to be where they really struggle, and I wonder if that is broken or perhaps it was never properly done at Alphabet.

I would say that Google has an even greater chance of launching a surprise product. For this reason, it is incredibly powerful. The counter-argument to what we were talking about here is that the search does not show, excuse me, the ads do not show a decline yet.

It is still showing growth. Therefore, this remains a growing and profitable company and is one of the strongest companies

I would say that the healthcare sector is perhaps where Google might really surprise us . If you consider the data that Fitbit possesses, in addition to its artificial intelligence algorithms , there may be a surprising moment coming in that field, and that will be of great importance not only to society, but also to business.

Okay, let's conclude here with our final thoughts on the stock, and I want to mention some of the assets that are on the balance sheet . But, as you know, Alphabet's price-to-earnings ratio as we record it today is 17 based on the past twelve months.

Incidentally, this includes estimated profits at their market value from some of the assets they own, particularly huge stakes in Anthropic and SpaceX. So, when you think of Alphabet, yes, think of the search engine, but also think of all these other businesses they own.

There's YouTube, and they have an Android business and everything related to it in terms of devices that they manufacture. So, some of this money even goes to support research activity, right ?

Because this is a great way they make money from Android. And you have all these other assets. They have a huge venture capital arm and a private equity arm . So, there are hundreds of billions of dollars worth of goods there.

The future price-to-earnings ratio may be a slightly better indicator of their valuation. It is currently around 26 . They also have very high growth in the cloud computing sector.

Do you believe that these combined assets make this company capable of adapting in the future? With that in mind, do you think valuation is something that should attract or scare investors at this stage ?

I think that's a good rating for this company. I don't think it's a very tempting deal, but I wouldn't worry about the price at which it's trading today. I think it's an incredibly strong company .

I would say that there is nothing wrong with keeping them within a diversified investment portfolio. It is a very strong company. It remains a very strong company even with some question marks.

I would say that this extraordinary possibility, at any moment and with all that supports it in the background, could unleash something very important. I would say that this is worth that chance of promotion.

I'm in roughly the same boat. I would say that I acquired my stake in Alphabet probably two years ago . With a very good rating , it was roughly in the middle of the dozen. As the price rose, I made a small profit at a P/ E ratio close to 30.

I've been thinking about selling some recently because I'm really concerned about the product side of the company. I think late last week the Gemini 4 model was introduced. Theoretically, I know they always tend to be more benchmark-oriented , so treat their benchmark data with extreme caution, but some things I've seen show that this might be a very good model.

I think one of the things we've learned with regard to the cloud is that many more companies are using Gemini than we probably realize because it's so cheap . Hmm, and that should be their advantage, shouldn't it?

They make the TPU units, they own the cloud, and they make the model itself. So, there is no ...they can keep all the profit margin internally and make everything run more efficiently, which is another thing they talked about regarding Gemini, which is actually: "Hey, this is very helpful in making us run the entire cloud more efficiently, and then we'll be able to have a cumulative cost advantage, maybe."

So, I keep going back to those things as a kind of non-panic response, but it's really on my list of concerns that research is probably under more pressure today than ever before.

This does not mean that they cannot answer those questions, but I will be watching closely as their quarterly results are released .

I have no doubt that the cloud will continue to grow at a really rapid rate. Perhaps that will simply outstrip the search over enough time. Then the product aspect is the other part that I really want to monitor.

Are they capable of making good products and taking advantage of the fact that my email is with Alphabet, my calendar, and that millions, you know, more than a billion people use Android devices?

So, as far as the operating system is concerned, they have all the advantages in the world. If they are unable to take advantage of that, then you will see that this will be a big problem for the company.

But I think we are somewhat in agreement that at the moment , holding shares may be an acceptable option, but it is not necessarily a very attractive buying opportunity. Therefore, keep Google on your watchlist for potential disruptions.