What the Chip Selloff Really Means for Investors
August 11, 2026 · Steve Laska, Matt Gomoll · 24 min

What the Chip Selloff Really Means for Investors
August 11, 2026 · 24 min
Episode summary
In this August 2026 episode of Plain Talk, Steve Laska and Matt Gomoll look at the sharp July pullback in semiconductor stocks and conclude it looks more like a rotation and a valuation reset than the end of the AI story. Matt explains that while chip stocks fell hard, the broader market stayed within a couple percent of all-time highs, because investors took profits in chips and moved money to other areas. He says the record profits at memory makers did not look sustainable for five years, both because new competitors should eventually enter and because the hyperscalers have a strong incentive to use chips more efficiently.
The hosts describe AI as many layers, from frontier model companies down to data centers, land, power generation equipment and chips, with value gathering wherever there is a bottleneck. Steve points to how long it takes to get a data center approved and built. Matt notes that Nvidia had trailed the S&P 500 over the past year and says it is healthy for other sectors such as financials and health care to lead for a while.
Matt compares the moment to the dot-com era, when the internet's impact was real but valuations ran ahead of earnings. Both hosts caution against knee-jerk exits and stress managing risk with a balanced, diversified portfolio.
Key takeaways
- Matt described the July chip selloff as a rotation in which investors took profits in semiconductors while the broader market stayed within a couple percent of all-time highs.
- Matt said record memory chip profits were unlikely to last five years, because new competitors should enter and AI companies have a strong incentive to use fewer chips.
- Matt explained that value in the AI build-out tends to gather at bottlenecks, including approved data center land and power generation equipment.
- Steve pointed to land, zoning, power, water and local lawsuits as reasons data centers can take years to bring online.
- Matt said Nvidia was up about 18% over the past year compared with just under 24% for the S&P 500 total return, and called it healthy for strong stocks to take a breather.
- Matt compared the moment to the dot-com era: the technology's impact can be real while valuations still run ahead of earnings.
- Both hosts cautioned against knee-jerk decisions to exit the market and stressed a balanced, diversified portfolio.
Questions this episode answers
Why did semiconductor stocks drop in July 2026?
In this August 2026 episode, Matt Gomoll described the drop as a rotation: investors who had paid a premium for chip stocks concluded that record profits were not sustainable for years and moved money elsewhere. He pointed to the likelihood of new competitors and to AI companies' incentive to use fewer, costly chips. He said the companies may still do well but that valuations had become stretched.
Does the chip selloff mean the AI boom is over?
Matt Gomoll said the selloff did not necessarily mean the AI thesis was wrong. He compared it to the dot-com era, when the internet's impact was real but valuations ran ahead of earnings. He expects that if AI proves economically useful, the benefits will spread to companies that use it, not just to the model makers, data center operators and chip suppliers.
Why do data centers take so long to build?
Steve Laska said the timeline starts long before construction: finding land near enough to the power grid and fresh water, getting it zoned, and then facing local lawsuits, which he said can push timelines out two or three years. Matt Gomoll added that this scarcity makes approved data center land and power generation equipment valuable parts of the AI supply chain.
Plain Talk reflects the views of the speakers as of the recording date and is provided for informational purposes only. It is not a recommendation to buy or sell any security or a guarantee of future results. Market conditions change; talk with your advisor before acting on anything discussed.
Full transcriptRead the full conversation
Automatically transcribed from the episode audio; may contain minor errors. Select a timestamp to play from that point.
Welcome to Trillium's Plain Talk. I'm Steve Laska, managing partner here at Trillium Wealth. And as usual, I'm here with my partner and chief investment officer, Matt Gomoll. Matt, I think we want to talk a little bit about, you know, we've had that AI euphoria and the chip trades and all of a sudden in July, it looked like, you know, I just saw the Odyssey and there's like a whirlpool just sucking everything down. Talk a little bit about what we saw in July and what we have to look forward to maybe. Yeah, it really was a rotation, right? And rotation means this is kind of trying to understand why part of the market is going down a lot collapsing in the chip sector. But the market wasn't, you know, was still within a couple percent of all time highs and that can be confusing. That can be like, you know, why is there all this volatility?
I'm seeing stuff on the news of these stocks being down and the AI trade being over yet the market held up. And it's due to the classic rotation effect where where you move from, you know, investors move from paying up for certain stocks, chips, you know, in particular were really hard hit. So they had been paying up a premium to own chip stocks. And then they say, Hey, I think this has gotten a little out over at skis, even though the earnings are insane, the revenue that that micron and other companies are producing is insane. They look and say, I don't think this is sustainable. I don't think this trend is a multi year trend, you know, forecasting out five years. And therefore, I'm going to take profits and move somewhere else. And that kind of precipitated the collapse in chip stocks.
Do we see that as long term though? Because, you know, I've talked about this before, you know, AI is going to move, there's a lot to AI. It's not small. It's a huge thing. And internally inside of AI, it's going to move at different speeds. Like, do we have enough data storage, you know, these data centers, do we have enough power to have enough water? Do we have all these things? But what we do know about chips is the more memory, the better AI works. Like that's, that's like a law of physics type of thing, right? It's just the law. Like if you have more memory, it works better. So are the prognosticators doing some profit taking now and saying we're waiting for the rest of AI to catch up? Or are they saying, like, there's other like, usually when chips start to fall or anything starts to fall, it's because there's there's new competition in there that's going to do it better, faster, cheaper, so to speak. What do we see in out there? Is it is it a, is it a wait for the rest of things to catch up? Or are there just new competitors in there?
Yeah, there's not there was a Chinese company that that you know, got listed in China and was all of a sudden one of the biggest stocks in their their market. This is one of the area that there shouldn't be these excess profits accruing right in a stable market. There shouldn't be Samsung and Micron and SK Hynix capturing these just massive profits. But the demand is so high and there is no other competition yet, right? So so I would say like, it's it's two fold one. There should be some new entrants into this space. Even though it might take a while, you just would say, I don't think it's sustainable for these just massive year over year growth numbers for these companies to persist. They should get competed away, even though it might take time. And then the other part is because these things, these chips cost so much money. There's a huge incentive for the hyperscalers in AI models to make themselves more efficient to use less of it, right? And so it's kind of those two things saying, Hey, we've, we've pulled forward a lot of profits. We've been able to price these chips at outrageous amounts because the hyperscalers are just paying for them to put them into the data centers. We don't think that's going to continue at this trend, right?
These companies are going to keep doing great, but it's that change back to cyclical. They're going to move, have ebbs and flows like chips have always done. And that's kind of what I think we're seeing is just saying, this is these record profits are amazing. We don't think it's like five years sustainable. And they certainly, the valuations on these companies really did get stretched. Not to say they're not going to be good companies going forward, not to say they're bad investments right today. Just that in all at once, the market, you know, and investors kind of said, Hey, this sector is a little hot and let's put our money to work somewhere that we can get more value. So do you think any of it comes down to like, like we see we've seen this, you know, in our own experience, I can just read about it, like, like how long it takes to get a data center up and running, like forget the actual build. First you come in and somebody says, we think this piece of land works.
And as again, we have some of our own personal experience with not every piece of land even works. Like the power grid isn't close enough to deliver the power. There's not fresh water that again, these things eat. So you have the first thing where you have to identify a piece of land, get somebody to agree to sell it to you. Then comes zoning. And this is not like a two week process. This is, you know, we've seen and heard stories where like somebody wants to buy a piece of land and they want to close on X date and that data ends up pushing out two or three years and there's always danger of it not even happening because as soon as you get approval, zoning, then come to lawsuits, right? And the lawsuits from local media, you know, local people go, don't want that here. Here's the reason why. And then that stretches out is part of this in effect of just like like, like again, these data centers, like they're, they, there's a lot of people that want to build them, but there's not a lot of places like Maine literally outlawed them now. Like the state of Maine said, no more in our state. Can that have an effect on chips?
Yeah, it could in that there's plenty of demand to buy them, but you can't actually put them places because you didn't build the data center, right? So what we're seeing with this, the AI ecosystem is many, many, many layers of which like the frontier model, the Anthropic, the OpenAI kind of sits high up, right? Like they're the ones designing how the AI works, but then under the hood is the data centers and the chips that are in the data centers and the energy that powers the data centers and the land and all these things. And whenever there's a bottleneck in any one of those spots, all of a sudden you see value a crew there because it's these things are in demand and a bottleneck creates scarcity. So to your point, because it's becoming harder and harder to get a data center approved, the land is very valuable, right? If you can get a piece of land zoned and approved, that's going to be a really valuable commodity compared to another use for that land.
And then you get the power generation now, like the power generation utilities are regulated, and so they're not having a ton of profits at the back of consumers and data center companies, but the companies that make the power generation equipment, GE Vernova, hugely profitable, like they are, you know, somebody has to buy the turbine engine or the other stuff that goes into a power plant or the electrical grid and there's scarcity there and you're seeing a lot of value created there by those companies. And so at each one of these places, there's just so much investment, so much money flowing into get these things up and running that it's being captured in all these unique little areas. Chips was, you know, the most recent one where people said, hey, there's not enough chips, and these AI models are becoming more and more memory hungry, right? And you get back to early AI models.
They kind of, you just have one chat and they don't really remember anything. They forget everything. You just talked about five seconds later and you'd start over. And now you're creating these context windows of a lot of information pulled in from past conversations and they're getting smarter and better. But that's memory helping to do that, right? And so that's kind of the driving force this year of why chips were just a really hot, hot pick. I think it validates, you know, on one of the recent podcasts I talked about IPOs and, you know, how they're coming out. We talked about IPOs and how, you know, they're coming out and you already know who they are, right? And I said, if you already know who they are and there's all that heat, you know, your quick rise or that huge return that you think you can get, that's probably already happened. This, the chip trade to me validates like that for all of our clients.
Like, I was talking about Micron right before we came on and you got the guys at Goldman still saying, well, we targeted 2000. It's at 900 right now. It's one year return is 720 something percent. But God forbid he bought it on June 20th. Right. I mean, that's, and that's kind of what we're trying to tell clients all the time, you know, as far as wealth preservation doesn't mean you're conservative. It means, you know, you're preserving your wealth and growing it within the context of your kind of future and things that you and I have talked about on wealth transfer, all those things. But I guess the question everybody's going to ask now is, you know, and you seem to be alluding to like, maybe this is like a resetting of the, of this, you know, of these stocks where it's like, okay, it's going to settle in for a while. It doesn't don't expect that it was down 30% in July and it's going to be up 60% next month because there's clearly something here that the market is telling us that this isn't your kind of normal, you know, profit taking so to speak. Like this is, this is a real thing that's endemic to what analysts are seeing in the, in the industry. Like this is a reset, not a profit taken. Is that what you're subscribing to right now?
Yeah, I think it's just, it's showing that the valuation got really stretched and investors kind of agreed, Hey, this is not going to be a sustainable plan. And so it's reasonable, right? So when you're up a thousand percent to fall 30%, while from the high watermark feels really painful. And if you invested at the high watermark, not good. But in the context of a one year return, very reasonable, very reasonable for these things to get a little bit back down to earth. And it's okay for stocks to correct over time, right? To fall to go sideways and allow the market to catch up and allow their earnings to continue to grow so that they can then move higher. You know, I was just pulling up Nvidia. We all think of Nvidia as just a high flying, you know, one of the best.
Yeah, the platinum standard. Yeah. The second biggest company in the world today, I think behind Apple, is those two go back and forth. It's underperformed the market over the last one year. You wouldn't have thought that. It's up, you know, one year return, about 18% compared to the S&P total return about just under 24. So for Nvidia, that's relatively going sideways, right? And so that's, it's those kind of things that it's okay for a stock that did that well to take a breather to underperform a little while. Some of the Mag 7 have been underperforming the S&P, and that's okay, because it shows that there's health in other places, right? For, for in 2025, it really felt like the Mag 7 were just driving everything and the rest of the, you know, other 500. Other 500 stocks were just kind of along for the ride and getting dragged.
And now it feels like there's some parts of the market, financials, health care that some traditional value sectors that are doing good. And we want that. That's, that's healthy for a bull market to say, Hey, some other, some other sectors can do well for a while. We don't have to have these face melting returns in chip and AI related stocks. But it's good for other parts of the market to have some returns and then lead for a little bit. Well, other tech things have done well. Like I thought I saw something more like Microsoft the last month or so is like the best performance since 1999. You know, Apple suddenly re-emerged as the most valuable stock in the world, right? I mean, yeah. So it is, it's even in the tech sector where other things have been like, we're not just kind of a one trick pony. But some other things going on here and like for a company the size of Microsoft to go up 25% in the last month or so is enormous. Right? Cause I mean, that's something that's followed pretty closely.
Like that kind of stock and Apple, these are not stocks anymore on these days that are like, you know, you're speculating on what they can't do. They're behemoths. So if they're moving like that, they're moving on real data, like real financials, real information. Yeah. And what's so interesting about especially these big ones, they have so many business lines, Microsoft, Apple, Amazon, Google, Amazon, they have so many different business lines that they're not, you know, in itself, it's a diversified business. It's not, I won't, I won't go so far to say it's diversified in the way that Berkshire Hathaway is a, is almost a pseudo index. It is. Yeah. There's a lot going on under the hood of a lot of different tech plays at both Amazon and Google. And Apple has done fantastic because it, it doesn't have all the CapEx, you know, capital expenditures that the other hyperscalers do. It's just said, Hey, we're going to sit back, let you guys play. We're going to make great hardware. And we believe AI is going to be on device and we're great for privacy. We're great for all these things.
And the App Store charges basically a toll, right? It's a, it's a great business model of we create a good device that everybody's in the world is going to have at their fingertips. And if you want to be on it, you're just going to pay us a, you know, a cut of revenue that you receive. Yeah, it's pretty brilliant. Actually, like I said, they're doing, they're doing really, really well. And it used to be that their stock was rising on, you know, the new iPhone coming out like nobody's even following that anymore. Like I had to look because I was like, at my phone's been a little slow. Should I get a new one or should I wait? It's like a new iPhone coming out in September and everybody's like, really? Like you have to, you have to literally Google to find out that that's happening. Whereas five, six, seven years ago, you know, we'd be talking about the next phone. For the last 90 days and people be standing in line for it. So these, these companies have done wonderful work and they've got wonderful leadership in them. Anything on this other than Matt, like it's like, Hey, this is just kind of the nature of the beast. And it's why you diversify. And it's why you don't get, you know, it's why you buy all these companies are great companies, by the way, like you said, I mean, in videos.
I don't want to jinx it. They ain't going anywhere anytime soon. I mean, yeah, it's there. It's powerful. It's got a great future. It's just not, you know, a sit back and respect, you know, and expect to get a thousand percent return in the next two years. Yeah, it is. I think it's healthy when the market corrects, you know, especially in high flying sectors and allows things to catch up and allows, you know, in investing, you don't want the narrative to get too far out in front. I mean, if we, if we take this back to the dot com era, those companies were absolutely right about the transformative effect of the internet without a doubt, right? Fast forward 15 years from 1999 and the internet had transformed society, commerce, all these things. It's just they were way too far ahead on the valuation based on the narrative, right? The earnings were not there for those companies at that time.
So here it's good to have pauses in the bull market for the earnings to start to catch up, right? We haven't yet seen all the productivity gains and all the profit gains that use of AI will create for the entire economy. So it's good to kind of reset things and not get too far out of our skis with any one of these sectors and just allow that the whole narrative to catch up. So I think that's a healthy thing. I do think over the next few years, if AI is going to be proven to be useful and economically valuable, the benefit will diffuse across companies, not just be concentrated in who makes the model. Who runs the data center, who hosts it, and who provides power to it. It's going to actually come to the companies who are saying, Hey, I just use it for my payroll or my accounting or my sales process or whatever it is that you're using it for that makes you better and more profitable. And that will be a good thing as that starts to catch up.
Yeah, that's where I'm at on this. I actually think that this is a very good sign when we talk about like of alleviating people's fears of, you know, AI is taking over the world the next year or two. I think this reset is basically showing that other things have to catch up and what are for it to be functional and are for it to be usable. I think a lot of people have figured out that it's not free. It's far from free. Right to actually use it. The more the more you get out there and use it and you get behind, you know, a lot of folks will be out there using a free version of AI and think that that's kind of it. And they're just using it as a search engine, right, or doing some, you know, doing an email as you start to use it like I think I'm starting to use it and others are right now like it's, it's not free. It costs and, and that will temper people as well. So I'm very excited by it. I think it's a great thing that like you said, it's getting diversified. The, you know, those seven stocks, it didn't just seem like it. I mean, they were driving the market. Right. And there's always, you know, kind of a reset. And I think that's good. It's good for the markets and it's real healthy. I'm, there's so much going on in the world right now that it's always shocking to like look at the market and say it's still going up.
Right. It's like, you know, it's still going up. There's wars going on. There's inflation going on. And it's still just kind of cooking along. And it's cooking along now without it just being these kind of like, you know, AI stocks, chip stocks, everything's kind of propped up there pretty good. So, yeah, well, I think that's for the future. And what you said there is a good reason that you. You never want to make knee jerk reactions and get out because of one thing, right? Inflation could have been enough reason to say, I don't trust this. Like, you know, I'm scared of inflation and therefore I'm scared of rates and I should be out of the market or this war went on oil prices. I'm worried about. I should be out of the market or look at this chip collapse. I should be out of the market. Like, there's always going to be some, some bad news to say, I don't trust this. I should get out and you have to overcome that. And the way you do it is with managing your risk, have a balanced portfolio. And whether through those because there's going to be, there's going to be tailwinds pushing things higher that you don't want to miss out on.
Yeah, I think the net answer is, and again, I'll be pretty emphatic on this one is it's get professional advice and use it. I mean, at the end of the day, we've talked about it before Vanguard talks about a 3.5% annualized difference between using professional advice and not. And that's one of the great direct companies of all time, right? It shouldn't be confused. Like, I think people look at something like that and go, well, I didn't get out performed by 3.5% this year or the next year or the next year or the next year. No, it's the 30 or 40% mistake you make. And like, get advice, take it. And like, here's an advertisement for good advisors. Like, go to advisors that are willing to push back at you on your ideas. That's the conversation I had with again with a teammates father yesterday. His wife had sold a house 20 years ago. And she has $150,000 from the proceeds that she put, she put a hundred. 20 years ago in a CD. I'm like, if she were an S&P fund, they would be worth $800,000. There was silence on the other end of the phone.
And I did say to them, like, I'm not saying your advisor isn't competent. They are. I'm saying they're not extraordinary. They're not extraordinary because their job is to not let you make that decision. So, this is a good example of where advice matters is getting out there and saying, because there's a lot of moving parts. And like you said, I think everybody thinks Nvidia is still cooking along. And I wouldn't sell it. But like you said, it's outperformed the broad market over the last 12 months. If you ask the average person on the street who knows who Nvidia is, what do you think Nvidia has done in the last year versus the US market? I would be shocked if they didn't say double, triple, quadruple. Right. So, it's a reason to get out there and listen and things are always moving. All right. On that note, we'll close today and look forward to our next ones. I don't know if we'll get it by the next time, Matt, but I do want to start to get some guests on here to talk about some different topics. And I'm looking at somebody to come on and talk about things like clean energy. I'm looking on somebody to come on and talk about the small business environment in the country today, where you and I reside.
You know, we run a small business today. And I think it's a wonderful place to be and want to encourage others to do that as well. So with that, Matt, as usual, thanks and look forward to next time. All right. Take care.
Keep listening.
The Real Costs and Benefits of AI Data Centers
October 6, 2026 · 30 min
Episode notes →TechnologyA New Fed Chair Nominee and AI's Disruption of the Software Sector
February 12, 2026 · 32 min
Episode notes →TechnologyIs AI in a Bubble? Separating Hype from Investment Reality
December 4, 2025 · 33 min
Episode notes →Start with a conversation.
Tell us what is on your mind. We will match you with an advisor near you, or meet by video.