AMD是强力买入，得益于Helios货架需求及AI合作伙伴关系；预计股价将在2-3年内超过600-700美元。

If you are an AMD shareholder, you will want to watch this video as we will be taking a close look at many of the things that have affected AMD in the past few days . Welcome back to another episode here about AMD. I think we haven't talked about this company for a few weeks. Overall, I remain very optimistic about this company. I still believe it's a player in a market worth trillions of dollars because the growth opportunity for Helios shelves will indeed pay off . And I truly believe that this is the reality, don't you?

For me, the big positive would be the "Helios" shelves. Unfortunately, the downside is that if there is no real demand for "Helios", then this company is overvalued . This is what I believe every investor should understand when investing in AMD.

I believe the demand exists. We have some great partnerships with companies like OpenAI, partnerships with Anthropic, with Microsoft, with Meta, with Oracle , and the list goes on and on . I think we will get more and more as time goes on.

But currently, the stock price is hovering around $450. Since the beginning of the year, the stock has risen by about 100%, so that's a very good return for investors. Unfortunately, the stock has fallen by about 21% from its 52-week high.

Fortunately, there are Broadcom and AMD. Both of these companies have secure supply chains. We know that AMD secures its manufacturing at TSMC. They are working on the entire supply chain because they understand the scale of this issue. Most likely, they are not expanding their reach because of their balance sheet as Nvidia is. But they control part of the supply. They also make supply deals to ensure that part of the supply chain is secured as well.

For me, this was one of the main reasons for feeling excited about AMD. Again, it all really depends on the AMD team, doesn't it? To check their engineering, I think they have a great engineering team. They have an excellent team in advanced packaging, but it's all up to them to ensure that the "Helios" system matches the specifications and speed that all these companies really need. I'm not just talking about performance, but about how this infrastructure is built , aren't I?

No AI company would want to buy a chip if it would take years to get it up and running, would it? Or it will take many, many quarterly periods to prepare it. This will not work , this will not be acceptable in this field. At that moment, they will say, " Maybe I can just wait for Nvidia to get more supplies, because by the time I get them, I know I can run their system at lightning speed." Therefore, AMD needs to execute extremely well during the Helios production ramp-up system, and I believe they are doing that. But this is one of the risks.

Now let's get back to AMD, okay? Here we continue the AMD story. The AI story is here for AMD, and again it all depends on Helios, but if Helios performs as strongly as I think it does, then AMD’s AI story has already begun . What do I mean by that ? What I mean by that is that we will continue to see consecutive revenue growth for AMD , right? It started off a bit in June , but I saw a slight stumble in March 2026, and now it's really starting to recover. But I believe that from now on , as long as there are no obstacles, the revenue growth story will continue to climb consecutively.

For the next quarter, we expect year-on-year growth of 40% to reach approximately 13 billion, and then this will accelerate in December 2026 to reach 16 billion with year-on-year growth of 56.7% . I believe the following quarter will also see an even greater acceleration in growth. I mean, if we look at AMD's position in December 2026, the market is expecting 46% year-over-year growth, and even after this strong base, it's expecting growth of approximately 72% in the following year. Therefore, AMD is definitely an integral part of this AI story.

Another thing I would like to highlight is the price-to-earnings-performing- futures ratio. The next twelve months look a bit expensive, at a level of 41.1. But in fiscal year 2027, that is, during the next 18 months, the ratio will drop to around 30. That doesn't seem crazy for a company that's just starting out in the AI story, does it? It's not even a billion- dollar player, and it has tremendous opportunities as these companies' capital expenditures continue to grow.

Firstly, Humane now offers productivity AI computing services in Saudi Arabia powered by AMD MI355 and AMD EPYC processors. True, the 355 model is not a rack-scale system . So, there is limited demand there, but it's great to see this progress.

Now, what I considered more important was that AMD stated that Cisco and Humane plan to deploy up to 250 megawatts of AI infrastructure as part of the next phase of expansion starting in 2027. The word "up to" is the key word here, but it illustrates that a few hundred megawatts of computing means a few billion dollars in revenue. One gigawatt is worth tens of billions, and I think it's closer to $20 to $30 billion in revenue for AMD. I mean, a quarter of that is still about $5, $6, or $7 billion. This is still a large sum from a single client, Humane, who is based abroad. They are still on track to deploy around 1 gigawatt of AI infrastructure by 2030. So, it's good to see this expansion . It's good that they're still planning to release MI450 processors in 2027 as well.

The other thing we saw was that AMD’s Instinct graphics processing units and Epic central processing units will power the next generation of the Lumi AI supercomputer in Europe. They will be using AMD's Instinct MI430 graphics processing units and AMD's sixth-generation Epic CPUs . Now, regarding the MI430, one of the things I love about AMD is their great prowess in advanced packaging. So, they are really able to use this microchip architecture and design the chip the way they need it, right ? You have MI450 processors for real AI accelerators and AI labs. As for sovereign AI, where you might need a mix of simulation and certain things that don't necessarily always focus on AI workloads , the MI430 is great for that. I think they have other solutions too, but this really allows AMD to enter that sovereign market, which is also growing like crazy. Therefore , this will be built on the MI430X processors and AMD's new CPU solution. Deployment is scheduled to begin in the second half of 2027. This means that some 430 processors will likely be sold in the first half of 2027 if deployment is planned for that time. So, these are more small customers, small customers here and there, starting to appear as one huge customer , right? If you start getting a lot of smaller players expanding through this infrastructure bet, and through all these different GPUs, that helps build that revenue. So, again, young players are good, especially when you keep seeing a lot of news like this.

The next thing we want to see is that AMD has changed the data center accounts. Dan McNamara, senior vice president and general manager of enterprise computing and artificial intelligence, wrote a blog post about it. He is essentially saying that there should be integration into AI servers now. And servers in general. Companies currently tend to have a lot of outdated products. These are some of the accounts that have been viewed previously. Before artificial intelligence , we had already seen this. Perhaps not to that extent, but we have witnessed high-level discussions. What we see here is that organizations can consolidate 100 servers running five- to seven-year-old Intel Xeon Platinum processors into just 14 servers running AMD EPYC processors. This represents an 86% improvement , with a 69% reduction in energy consumption and a 41% reduction in the total cost of ownership over three years. So, once again, they were presenting this as a real reason to push in the direction of: "Hey, you need to upgrade those old processors, which are 5 or 6 years old, because you'll save a lot of money ." But now, the big difference is that you are limited by the amount of energy you have available. So, if you can reduce the amount of power consumed by these processors in all these servers, you gain space and, more importantly, you gain power that you didn't have before. And now, thanks to this power, you can say: "Look, I can now buy a few GPU servers or AI servers for my organization." The bottom line is that server consolidation may not be the most prominent item on IT officials' agenda right now, but it is a tangible and measurable project. So, eventually, maybe not this quarter or next, but there will come a time when IT officials will say, "It's finally time to upgrade those old server processors." It's time to add new server processors to get better space and, more importantly , it opens up the possibility of buying some AI servers as well. So I think it's interesting to realize that there is this kind of potential additional support for AMD.

Then, on August 27, just over a week ago, we saw AMD launch ROCm 10. This is their software platform to help work with all of their solutions. It's great to continue seeing these updates. They stated that this is a new AI-powered development platform, and they added many AI tools and AI solutions to help with software development as well. So it's great to see AMD not only focusing on rack systems, individual processors, or individual graphics processing units, but also focusing on the software domain. With artificial intelligence, this process is accelerated at rapid levels.

Now, on Friday morning, I think around 5:00, AMD gave what looked like a keynote address. I think she was in Europe, at the IFA 2026 conference . We used to talk a lot about local artificial intelligence devices there. AMD has put forward some interesting solutions with their business partners. We see the ThinkCenter device from Lenovo. We are now entering a market where we also need some local artificial intelligence . AMD has already done that. I think they showcased the AI Halo technology during CES or maybe before that. Now, we're moving to more similar solutions, and you'll be able to connect up to four of these devices to boost performance and get powerful local AI solutions. They actually presented concepts for laptops.

So, it seems that this is the direction the market is heading because AMD has now started doing it as well. I think one of the main drivers here was Apple. Apple was capturing a lot of market share, wasn't it? MacBooks were selling , and Mac Studios computers continued to sell. AMD and Nvidia say this is a market opportunity and we need to get into it with this AI story. This also allows them to continue expanding their software system , doesn't it? AMD is able to expand its software ecosystem around itself to attract potential new developers in the future.

So, in my opinion, there's a lot of great news here for AMD. I don't know where the stock price will be this year or where it will end up, but I still strongly believe that this stock will reach a price of more than $600, and even $700, within the next two to three years . In the