AI Revolution Goes Beyond NVIDIA to Semiconductor Component Manufacturing Stocks

June 4, 2025

The AI revolution has seen investor returns skyrocket for certain semiconductor manufacturing companies.

AI sector is the focus in this interview of Tore Svanberg, a Wall Street Journal’s “Best on the Street” Analyst and Managing Director at Stifel Financial Corp.

Tore Svanberg, Analyst, Managing Director, Stifel Financial Corp.

Tore Svanberg is an Analyst and Managing Director at Stifel Financial Corp.

Mr. Svanberg is a Managing Director in the Technology group, covering semiconductors with a focus on analog, connectivity and processor semiconductors.

He has been an analyst for 22 years and has been recognized for his work by The Wall Street Journal’s “Best on the Street” Analyst Survey.

His insight into the upstream components of the AI expansion for semiconductor profits is an important consideration for investors looking to capitalize on this trend.

“First of all, you’re right; there is an exemption for semiconductors, especially exports coming out of Taiwan.

As you may know, Taiwan is the hub, especially for fabless semiconductor companies where you have TSMC (NYSE: TSM) manufacturing wafers for U.S. semiconductor companies.

So, semiconductors are exempt today.

Again, that would be that direct impact, right?

Then, if you think about moving further up the value chain to actual phones and other consumer electronics, yes, you’re right, some of those are also exempt.

But again, keep in mind that as of today, there’s basically a blanket 10% tariff on anything.

Then there’s exemptions.

Obviously, China is the big topic, where the tariff rate is significantly higher than that 10%.

And I guess that’s where, obviously, there has been some encouraging developments from the negotiations that get started on Saturday.

Both parties are working to bring tariffs down, but there is still some uncertainty as it relates to the final numbers as the negotiations include other topics than just economics and trade.

But what’s very interesting, again, from an electronics perspective, is that this sort of U.S.-China trade war didn’t start on “Liberation Day.”

It started almost a decade ago.

I would say a lot of electronics manufacturing has already moved outside of China, especially to places in Southeast Asia.

So here we have countries like Malaysia, Vietnam, Cambodia, and, as you may recall, on April 2, some of those tariffs were very, very high.

I think that’s why investors kind of freaked out, because if those tariffs are going to be high, we have a serious problem.

Obviously, now we have this 90-day window where those tariffs are now more at 10%.

And then there are negotiations going on with each country to see where that number is going to settle down.

But yeah, already a lot of manufacturing has moved away from China and into Southeast Asia.

I think that’s a very important point to make.”

The AI computing process in very energy intensive and requires specification compliant components to make it all work.

Certainly AI has become a very important growth driver for the semiconductor industry.

But I think what’s sometimes not clear from an investor perspective is how AI is going to impact the whole space.

Up until now, of course, everyone’s been really excited, but Nvidia (NASDAQ:NVDA), because they are the leader in GPUs and obviously GPUs are the bedrock of high-performance compute that you need for AI.

But I think what’s not as well understood is that it’s not just about compute.

You need networking technology, you need power.

There are a lot of semiconductors that go into building an AI cluster.

The two areas that I’m very focused on are on the networking side.

There, you have copper connections and optical connections, and then there are all different types of standards within the networking industry.

And then on the power side, of course, a lot of these AI clusters use an immense amount of power, all the way from the grid to the actual GPUs and CPUs themselves.

There are certainly companies there that I cover that benefit from that as well…

On the networking side, we have companies like Marvell (NASDAQ:MRVL).

They are benefiting from the optical networking that is required in these AI clusters, especially for scale out of AI.

Then, on the copper side of things, which is used more for scale up, we have companies like Credo Technology (NASDAQ:CRDO), also Astera Labs (NASDAQ:ALAB).

All three of these companies have very high exposure to AI infrastructure.

If you then take it a level further down, you have companies like MACOM (NASDAQ:MTSI) and Semtech (NASDAQ:SMTC), which are also benefiting quite a bit from both optical and copper connections.

But their percentage of exposure to these markets is not as high.

So, here we’re talking about maybe 20%, 30% exposure, whereas the previous three companies that I mentioned have anywhere between 60% and 100% exposure to AI infrastructure.

On the power side of things, we really like Monolithic Power Systems (NASDAQ:MPWR).

This is a company that historically has done really well handling the power for the processors themselves.

Whether it’s GPU or CPU power management, they are the leader.

But they’re now also taking things to the next level and working on what we call rack power.

They’re trying to optimize the entire power envelope all the way from direct current transmission that’s coming into the data center, all the way to the actual processor power management.

Monolithic Power is definitely a company that I think is going to continue to benefit a lot from the trends we’re seeing for rack power.”

The AI sector growth may also lead to even more semiconductor stock upside.

“Companies like Texas InstrumentsOnsemiNXP, they all talked about the beginnings of a recovery in the industrial market, which was a bit of a surprise to investors.

I still think there’s a debate out there, whether that is tied to pull-ins because of tariffs or whether we are actually starting to see the beginning of a new capex cycle.

My guess is that it’s probably a little bit of both of those things.

Obviously, if we’re going to see a more sustainable capex cycle, we are going to have to see interest rates continue to come down.

That’s very important for industrial capex.

But we also obviously need to see a little bit more certainty around the tariffs and how those play out over the next few months.

But I would say that is the one area that was a bit of a surprise so far during earnings season.”