Speakers argue AI infrastructure demand remains strong despite political volatility and valuation compression in semiconductors.
Summary
Speakers analyze the market divergence following Democratic election gains, noting a sell-off in semiconductor suppliers despite strong earnings and historically low forward P/E multiples of 13x-14x. They contrast the current AI cycle with the dot-com crash, emphasizing that AI is demand-driven rather than supply-driven. The hosts highlight specific opportunities in TSMC, Nvidia, and robotics stocks like Ouster, while dismissing fears of a capital crunch for major hyperscalers. They conclude that long-term AI model capabilities remain on trend despite short-term political and financing noise.
Episode sections
- 0:27 – 5:48
Political Impact on AI Market Dynamics
Speakers discuss the immediate market reaction to Democratic gains in the House and Senate, noting a divergence where broader indices rose while semiconductors fell. They analyze how political uncertainty regarding data center regulations, particularly in Texas, has triggered a sell-off in infrastructure suppliers despite strong earnings from hyperscalers.
- 5:48 – 11:40
Earnings Review and Supplier Performance
The hosts review recent earnings from AI supply chain companies, highlighting strong results from Air Test Systems, Lattice, Edison, Advanced Energy, Fabrinet, Arista Networks, AMD, Navitas, Coherent, and Ouster. They note that despite bullish financial data and raised guidance, these stocks are selling off due to broader market sentiment and political fears.
- 11:40 – 20:34
Market Valuation and Historical Context
The discussion shifts to market valuations, with semiconductors trading at historically low forward P/E multiples (13x-14x) despite high expected growth. The speakers compare the current sell-off to the dot-com crash and 2022 bear market, arguing that the AI cycle is demand-driven unlike the dot-com era, which had excess capacity and no demand.
- 20:34 – 29:44
Capital Constraints and Future Demand Drivers
Speakers address concerns about capital constraints for 'NeoClouds' like OpenAI, citing Elon Musk's warnings and reduced fundraising targets. However, they counter this by pointing to massive capital raises by hyperscalers (Meta, Microsoft) and other players (SpaceX, Anthropic). They conclude by discussing future demand drivers, specifically AI applications in knowledge work and life sciences.
- 29:44 – 31:06
Agentic AI and R&D Applications
Speakers discuss the nascent stage of agentic AI applications in knowledge work and its transformative potential in R&D sectors like mathematics, material sciences, and life sciences, noting regulatory constraints in the latter.
- 31:06 – 33:44
Portfolio Concentration and Demand Risks
The discussion shifts to portfolio management, highlighting that semiconductor, memory, and optic stocks are trading based on fears of 2027 demand drops due to political or credit tightening, which could erode pricing power and peak margins.
- 33:44 – 38:45
Compute Bucket Analysis: TSMC, Nvidia, Broadcom, Intel
Speakers evaluate compute stocks, identifying TSMC as a safe anchor with strong profit expectations, Nvidia as having moderate risk due to neo-cloud financing exposure, Broadcom as higher risk, and Intel as a high-risk/reward turnaround play dependent on foundry execution.
- 38:45 – 45:54
Memory and Optics Cyclical Dynamics
The memory and optics sectors are described as highly cyclical with similar supply/demand dynamics. Speakers note Marvell's acquisition of Lumentum as a move toward vertical integration in interconnects, while expressing caution about ACM Research's China exposure.
- 45:54 – 49:45
Robotics Sector Valuation and Growth
Robotics is identified as a key area for portfolio expansion due to improving valuations and advancements in foundation models for robotics. Speakers expect the robotics allocation to at least double, citing companies like Regal Rex Nord for stability and Ouster for high growth potential.
- 49:45 – 53:45
Skilled Labor Stocks Post-Election Reaction
Skilled labor stocks like IES Holdings and MCore experienced significant post-election declines. Speakers argue this is an overreaction, citing chronic labor shortages and universal grid upgrade needs as factors that should support pricing power despite potential data center slowdowns.
- 53:45 – 1:00:11
Semiconductor Equipment and Ultra Clean Holdings
The final segment covers semiconductor equipment, noting that extreme growth rates cause volatility. Ultra Clean Holdings is highlighted as trading at a deep discount (7x 2027 earnings) after a 70% drop, presenting a potential long-term opportunity if core AI demand remains intact.
- 1:00:11 – 1:01:09
Valuation Opportunities and Portfolio Review
Eric states that AI investment opportunities are better on a valuation basis than six months prior, despite current portfolio losses. He notes that the underlying trends driving initial valuations remain intact. The host asks for parting thoughts regarding recent unusual losses in the market.
- 1:01:09 – 1:03:02
Market Volatility and Priors Reassessment
Eric describes recent market drops as brutal, citing 'Deep Seek' and the 'tariff tantrum' as gut-check moments that forced a re-evaluation of investment priors. He admits to being too blasé in previous episodes regarding downstream financing implications and spent 20 hours analyzing his portfolio to understand stock movements.
- 1:03:02 – 1:07:24
Long-Term Demand Thesis
Eric identifies the core question for AI investing as whether demand will stay elevated in 2027 and 2028. He asserts that AI model capabilities have not deviated from their previous trend lines despite market volatility, which he finds heartening. He suggests staying the course could be richly rewarded if demand holds.
Assets discussed
| Asset | Kind | Stance | Context |
|---|---|---|---|
| AmazonAMZN | equity | mixed | Amazon, Microsoft, Alphabet, I believe each of them was up more than 5%. |
| MicrosoftMSFT | equity | mixed | Amazon, Microsoft, Alphabet, I believe each of them was up more than 5%. |
| AlphabetGOOGL | equity | mixed | Amazon, Microsoft, Alphabet, I believe each of them was up more than 5%. |
| NVIDIANVDA | equity | mixed | So, you know, it's it's one of our listeners, Mike McMahon, he had captured it on um Twitter pretty well, and I I think he was observing this about NVIDIA, but you you look at all this data and you say this is either the fattest pitch that AI bulls have seen within the entire cycle, or the market's gonna look make make me look really stupid a year from now. |
| Meta PlatformsMETA | equity | neutral | We had Meta announce plans to raise up to $97.5 billion of equity. |
| Broadcom Inc.AVGO | equity | neutral | Taiwan Semi, Broadcom, Nvidia, Intel all go in that compute bucket. |
| COHERENT CORP.COHR | equity | neutral | If you go back and you look at the share charts for a company like Coherent, um, when when when oversupply situations come, they can be incredibly brutal to both of these industries. |
“Stance” is how the speakers talked about the asset, not a recommendation.
Claims worth verifying
The S&P 500 was up roughly 2% and the Dow up roughly 3% on election day, while the semiconductor index was down 4%.
Eric ·0:46
How to verify: Check historical market data for the S&P 500, Dow Jones, and SOX index on the release date 2026-11-05.
Amazon, Microsoft, and Alphabet were each up more than 5% on election day.
Eric ·1:19
How to verify: Check historical stock price data for AMZN, MSFT, and GOOGL on 2026-11-05.
Prediction markets had an 80% probability of Democrats retaking the Senate and House before the election.
Eric ·1:24
How to verify: Review historical prediction market data (e.g., PredictIt, Kalshi) for the 2026 US midterm elections.
Texas has approximately 33% of data centers under construction.
Eric ·1:43
How to verify: Review industry reports on data center construction pipelines in Texas (e.g., from JLL, CBRE, or Uptime Institute).
Greg Abbott lost the Texas gubernatorial election.
Eric ·1:51
How to verify: Check official election results for the 2026 Texas gubernatorial race.
43 out of 44 ballots regarding data centers passed in Texas elections.
Eric ·2:04
How to verify: Review official ballot measure results from the 2026 Texas general election.
Air Test Systems stock rose approximately 30% the day after earnings and raised guidance to $150 million.
Eric ·6:32
How to verify: Check Air Test Systems (ATRS) stock price history and earnings press releases for the relevant quarter.
Lattice revenue grew 118% and guided for at least 60% growth in 2027.
Eric ·7:51
How to verify: Check Lattice Semiconductor (LSCC) earnings reports and guidance announcements.
Edison's backlog increased by 46% and track backlog by 59 gigawatts in the last quarter.
Eric ·8:13
How to verify: Check Edison International (ED) earnings reports and investor presentations.
Advanced Energy saw a 34% sequential increase in data center revenue.
Eric ·8:45
How to verify: Check Advanced Energy Industries (AEIS) earnings reports for sequential revenue breakdowns.
Agentic AI is less than a year old in terms of its broad applications.
Unidentified speaker ·29:44
How to verify: Check industry reports on the launch dates of major agentic AI platforms and enterprise adoption metrics.
AI has already been transformative in mathematics and is expanding into material sciences and life sciences.
Unidentified speaker ·30:30
How to verify: Review academic publications and industry case studies on AI applications in mathematical proofs and material discovery.
TSMC is expected to do more than $150 billion in net profits by 2028.
Unidentified speaker ·34:26
How to verify: Check TSMC's financial filings, analyst consensus estimates, and company guidance for 2028 profit projections.
TSMC has a $1.8 trillion market cap.
Unidentified speaker ·34:32
How to verify: Check real-time financial data for TSMC's market capitalization.
Nvidia has expectations for $16 in EPS next year.
Unidentified speaker ·35:42
How to verify: Check analyst consensus estimates for Nvidia's earnings per share for the upcoming fiscal year.
Intel lost 60% of its value due to a misstep in allocation between the data center/AI market and PC market.
Unidentified speaker ·36:44
How to verify: Review Intel's stock price history and earnings call transcripts discussing strategic shifts and market performance.
Micron was trading at $80 a share when first recommended and is now at $800 a share.
Unidentified speaker ·39:55
How to verify: Check Micron's stock price history at the time of the podcast's previous recommendations versus current price.
Marvell acquired Lumentum in an all-stock deal.
Unidentified speaker ·42:32
How to verify: Review the official merger agreement and press releases from Marvell Technology and Lumentum.
IES Holdings is down 17% and MCore is down 11% on the day of filming (Nov 4, 2026).
Unidentified speaker ·49:54
How to verify: Check historical stock price data for IESC and EME on November 4, 2026.
Ultra Clean Holdings is down 70% from its June high and trading at 7x its 2027 earnings.
Unidentified speaker ·56:08
How to verify: Check Ultra Clean Holdings' stock price history from June 2026 and current valuation multiples based on analyst estimates.
Opportunities on a valuation basis are much better than they were six months ago.
Eric ·1:00:11
How to verify: Compare current valuation multiples (P/E, P/S) of AI stocks against data from 6 months prior (May 2026) using financial data providers like Bloomberg or Yahoo Finance.
Eric spent 20 hours going through his portfolio to understand why stocks are moving and what he might be missing.
Eric ·1:02:40
How to verify: This is a personal anecdote and cannot be externally verified, but it serves as a claim about his recent activity level.
Amazon fell 90%, 60%, and 40% at various points in its history.
Host ·1:04:38
How to verify: Check historical stock price data for Amazon (AMZN) to identify maximum drawdowns from peak to trough during its growth phase.
The majority of stocks actually underperform the index.
Host ·1:05:54
How to verify: Review SPIVA (S&P Indices Versus Active) reports or similar academic studies on active vs. passive fund performance and individual stock returns relative to benchmarks like the S&P 500.
Allegro MicroSystems, IES Holdings, Onto Innovation, Power Solutions, Regal Rexnord, and Mcore are all down over 30% in the host's portfolio.
Host ·1:06:10
How to verify: Check the current price of these stocks (ALGM, IESC, ONTO, PSIX, REX, MCRB) against their purchase prices or recent highs to calculate the percentage decline.
Do your own research
- Verify forward P/E multiples for semiconductor ETFs and major suppliers using Bloomberg or Yahoo Finance.
- Review recent earnings reports and guidance from Air Test Systems, Lattice, Edison, Advanced Energy, Fabrinet, Arista Networks, AMD, Navitas, Coherent, and Ouster.
- Check capital expenditure announcements and fundraising data for Meta, Microsoft, SpaceX, and Anthropic to assess 'NeoCloud' financing health.
- Analyze TSMC's profit projections and market cap valuation relative to consumer staples peers.
- Assess the impact of Texas data center regulations on infrastructure suppliers like Ultra Clean Holdings.
- Review the acquisition details of Marvell's purchase of Lumentum for vertical integration signals.
- Monitor AI model capability trend lines from major labs to distinguish technological progress from market sentiment.
- Examine skilled labor supply data and grid upgrade requirements to validate pricing power arguments for labor stocks.
Full transcriptexpand
- 0:00
- Unidentified speaker: Eric, this is the episode the AI Investor Podcast listeners have been requesting for months.
- 0:04
- Unidentified speaker: This is where you are going to do a full review of your million-dollar AI portfolio.
- 0:09
- Unidentified speaker: We're going to start by looking at earnings and what we're seeing in the numbers, which stocks have won, which stocks have lost.
- 0:14
- Unidentified speaker: We're going to take a look at the big picture and tech valuations today and what pockets are attractive and which are unattractive or overpriced.
- 0:20
- Unidentified speaker: And then finally, we're going to get to the main event, which is the portfolio review and some of your top stocks today.
- 0:25
- Unidentified speaker: I'm very excited about this episode.
- 0:27
- Unidentified speaker: As we have said many times, this is not a political show, but we are on the eve of a pretty major political upset and the blue wave, which was promised, looks like it actually, you know, landed on shore, and there's sort of a changing of the guard in the House and the Senate.
- 0:40
- Unidentified speaker: Do we talk about politics as it affects the AI portfolio or do we just jump move on to the stocks?
- 0:45
- Unidentified speaker: What do you want to do here?
- 0:46
- Unidentified speaker: Well, I think we have to.
- 0:47
- Unidentified speaker: You look at the performance of the market today, and the Dow is up roughly 3%.
- 0:52
- Unidentified speaker: Excuse me, S P 500 closer to 2%.
- 0:56
- Unidentified speaker: And then the semiconductor index, the stocks, is down 4%.
- 0:59
- Unidentified speaker: So you have a six to seven percent divergence between stocks, um, you know, the broader economy stocks, and and and what's often concentrated into kind of AI investing.
- 1:12
- Unidentified speaker: So a couple things on this.
- 1:14
- Unidentified speaker: Number one, the largest hyperscalers are actually seeing phenomenal returns today.
- 1:19
- Unidentified speaker: Amazon, Microsoft, Alphabet, I believe each of them was up more than 5%.
- 1:24
- Unidentified speaker: And while it was consensus that there was an expectation that the Democrats would retake the Senate and the House, I think in prediction markets, it was around 80% headed into the election.
- 1:35
- Unidentified speaker: I think the scale of the victory is what is providing some of this kind of outsized reaction today.
- 1:43
- Unidentified speaker: You look at a place like Texas, which has something like uh 33% of data circus under construction.
- 1:49
- Unidentified speaker: That could change depending on how you look at the metric.
- 1:51
- Unidentified speaker: But um, the Republican governor there, who was expected to win Greg Abbott, he he lost.
- 1:57
- Unidentified speaker: And there's been talk about a moratorium there.
- 1:59
- Unidentified speaker: And there was, I believe, 44 ballots on data circus.
- 2:04
- Unidentified speaker: I think 43 of them passed.
- 2:07
- Unidentified speaker: Only one didn't in uh North Dakota.
- 2:09
- Unidentified speaker: So that kind of shows where the mood is on data sirens right now.
- 2:13
- Unidentified speaker: But this has led to an extreme sell-off in stocks that provide kind of the shells and the electrification for data centers.
- 2:20
- Unidentified speaker: Those are seeing the highest sell-offs.
- 2:22
- Unidentified speaker: But the chips inside data centers are also seeing a sell-off.
- 2:25
- Unidentified speaker: Well, we'll talk about the details more, but the concept here is that you're probably going to have a harder time building data centers, which is going to lead to less leverage for suppliers.
- 2:37
- Unidentified speaker: As we've talked about on this show, one of the concerns is that if you have a slowdown, it doesn't necessarily mean that all of a sudden your semiconductor companies, because the AI market's growing 20%, they are growing 20%.
- 2:50
- Unidentified speaker: A lot of them have elevated gross margins relative to what they've ever been able to capture.
- 2:55
- Unidentified speaker: And if you lose that pricing power, you're going to not be able to grow earnings in line with kind of market growth.
- 3:02
- Unidentified speaker: So the market as of today is reading this as basically more power to the largest companies, the Microsofts, the Amazons, because number one, they already have compute online.
- 3:14
- Unidentified speaker: If if there's a lot of demand for AI services and not much new supply, well, the companies that have existing supply, it's going to become even more valuable.
- 3:26
- Unidentified speaker: And also, there's the belief priced in by the markets that this is going to lead to kind of more, you know, you'd call it pricing power from the hyperscalers themselves to push back on a lot of the um chip companies themselves.
- 3:41
- Unidentified speaker: So this is this is part of a broader rotation that we've seen across the past month, but it's really providing an accelerant.
- 3:49
- Unidentified speaker: The only thing I'll note um before we move on, and we'll talk a lot more about this broader theme, is that you often see really outsized moves the day after elections.
- 3:58
- Unidentified speaker: I remember the day after Trump was elected, I think the SP 500 for a bit was down something close to 10%.
- 4:05
- Unidentified speaker: And then it went on a significant rally in the months that followed.
- 4:09
- Unidentified speaker: So I would I would be cautious about reading too much into knee-derk reactions.
- 4:14
- Unidentified speaker: There's still going to be a lot to be determined.
- 4:16
- Unidentified speaker: For example, the governor uh elect in Texas a lot of what they would want Dass Iris to do uh before ending a moratorium on new approvals is things that most projects are already aligned with.
- 4:29
- Unidentified speaker: So it might be that you talk a little tough on it, but how it impacts actual new supply into the market is kind of overstated in that initial reaction.
- 4:38
- Unidentified speaker: So um, yeah, I think that's that's kind of the big takeaways from what we saw last night and what we're seeing in the market this morning.
- 4:45
- Unidentified speaker: Yeah, I mean, I I'm I don't think you should ever read too much into the market's first reaction and your Trump example there is a good one.
- 4:52
- Unidentified speaker: This feels to me to be an overreaction.
- 4:54
- Unidentified speaker: I would be surprised if the Democrats pulled back on data center build-outs as much as the market is acting like they will, because it is now, it's almost in a too big to fail moment.
- 5:05
- Unidentified speaker: In fact, it is.
- 5:05
- Unidentified speaker: It's too big a part of the GDP, it's too big a part of the economic growth story.
- 5:09
- Unidentified speaker: Um, nonetheless, you know, lots of fireworks for whoever watched and probably some impacts on the portfolio, although let's not try and predict them here, you know, 12 hours after the election.
- 5:19
- Unidentified speaker: I do wonder if something like SpaceX, which has a lot of capacity already, to your point, um, and a demonstrated ability to bring behind the meter online with, you know, just somebody who, you know, through sheer force of will can get things done and understands engines and turbines and power delivery.
- 5:36
- Unidentified speaker: I wonder if SpaceX maybe their compute actually could be at a premium here, as they're more able to build out behind the meter and other standalone data centers are not.
- 5:44
- Unidentified speaker: But time will tell.
- 5:45
- Unidentified speaker: Time will tell.
- 5:46
- Unidentified speaker: Let's um enough speculation though.
- 5:48
- Unidentified speaker: I actually want to get into the numbers.
- 5:50
- Unidentified speaker: We've had a lot of earnings, and we talked about some of the earnings in our last episode, but then there was additional earnings that followed.
- 5:56
- Unidentified speaker: So I want to hear about air test systems in particular, one of the stocks in my portfolio.
- 6:01
- Unidentified speaker: I think it was up, I don't know, 15 or 20% after earnings.
- 6:03
- Unidentified speaker: So talk to me about some of the other companies that have reported since the last time we spoke and what you're seeing in their figures.
- 6:10
- Unidentified speaker: Yeah, it's interesting because we we've once again seen uh a large sell-off.
- 6:15
- Unidentified speaker: We we will see one today concentrate primarily in semiconductors and chips and and suppliers into data centers.
- 6:23
- Unidentified speaker: But this is all coming at a time.
- 6:25
- Unidentified speaker: Earnings continue to be very good.
- 6:27
- Unidentified speaker: So I thought it would be worthwhile just before we got into some more details, just recapping what we've seen this week.
- 6:32
- Unidentified speaker: Um, as you noted, air test systems, uh, I think they were up actually closer to 30% the day after earnings.
- 6:40
- Unidentified speaker: So it was quite the reaction.
- 6:41
- Unidentified speaker: And what what they did was they moved guidance to the high end of their expectations, which is a $150 million this year.
- 6:49
- Unidentified speaker: But more importantly, they discussed several major customer wins uh that are coming closer to large orders.
- 6:57
- Unidentified speaker: So we've talked about burn-in test demand and the value proposition from a company like Air Test Systems and and how you're looking for differentiated companies, right?
- 7:06
- Unidentified speaker: There's there's a lot of companies out there with exposure to the biggest trends you want, but there are some companies with um kind of a better value proposition than others.
- 7:15
- Unidentified speaker: And and we've talked about advanced packaging and how complex chips are going to get in an AI world and and the test demands for that and and how an air fits into that.
- 7:25
- Unidentified speaker: So this is a company that it's going to be extremely volatile.
- 7:29
- Unidentified speaker: It has been extremely volatile.
- 7:30
- Unidentified speaker: And and if we see a continuing market sell-off, it's likely going to be sold with that.
- 7:35
- Unidentified speaker: But it's also a company that is giving us a read-through into kind of this bleeding edge of advanced packaging and and new optics applications, and they continue to win new customers.
- 7:46
- Unidentified speaker: Then we see uh Lattice, which is a field uh programmable gator company.
- 7:51
- Unidentified speaker: Revenue grew 118%.
- 7:53
- Unidentified speaker: They guided to at least 60% growth in 2027.
- 7:57
- Unidentified speaker: They issued an expectation of 30% annualized growth over the next five years and a backlog well into 2028.
- 8:05
- Unidentified speaker: We're also highlighting humanoid robotics as a meaningful contributor where they expect an inflection point the coming years.
- 8:13
- Unidentified speaker: Eden, which is a broad electrical system supplier, we've talked about their backlog in electrical rows 46%.
- 8:20
- Unidentified speaker: Uh, they their track backlog was up 59 gigawatts in the last quarter.
- 8:25
- Unidentified speaker: They they discussed runway beyond 2028 into 2029.
- 8:30
- Unidentified speaker: Um, so again, this is this is this is a company that issuing extremely strong results.
- 8:36
- Unidentified speaker: We have advanced energy, that's going to be read-through into the data center power market, which we've talked a lot about, and also semiconductor manufacturing.
- 8:45
- Unidentified speaker: They issued 34% sequential increase in their data center revenue.
- 8:50
- Unidentified speaker: So that's quarter over quarter.
- 8:51
- Unidentified speaker: Just in the span of one quarter, they saw 34% growth and they raised their 2027 expectations.
- 8:58
- Unidentified speaker: Fabronet, another company that's in the portfolio, they saw data center revenue jump 78%, talked about expanding capacity from being able to achieve $5.8 billion in revenue with a line of sight up to $13 billion.
- 9:11
- Unidentified speaker: So Austin, I went rapid fire there, but that was just all things that the market was ingesting on Tuesday.
- 9:19
- Unidentified speaker: And then on Wednesday, which is today, Arista Networks said that they expect their AI growth rate next year to be 100%.
- 9:27
- Unidentified speaker: Uh AMD had a very strong quarter with 27% sequential growth and data centers uh growing 50% sequentially, and they expect servers to double again in 2027.
- 9:40
- Unidentified speaker: Navitas, which is one of the more speculative power management plays, they talked about their AI infrastructure business expecting to triple in 2027.
- 9:49
- Unidentified speaker: That's before some of the big uh drivers with 800V get started in 2028.
- 9:54
- Unidentified speaker: Sterile Labs, which is a broad AI networking company, saw revenue growth of 154%, with guidance implying 27% sequential growth.
- 10:04
- Unidentified speaker: And Ousta, our our little robotics company that could.
- 10:08
- Unidentified speaker: A little LIDAR company that could.
- 10:10
- Unidentified speaker: I'm betting on you, Alster.
- 10:11
- Unidentified speaker: I know you've been a dog for me lately, but I'm betting on you.
- 10:14
- Unidentified speaker: I do like the tech.
- 10:15
- Unidentified speaker: Well, and they are bucking the trend today.
- 10:16
- Unidentified speaker: I could I could look at them as we record right here, but uh let's look.
- 10:21
- Unidentified speaker: They are up uh six percent.
- 10:22
- Unidentified speaker: They were up 14%.
- 10:27
- Unidentified speaker: Yeah, they were up 14% earlier.
- 10:29
- Unidentified speaker: And they they talked about their Rev8 uh LIDAR growing from hundreds of units in in Q3 to tens of thousands of units in in the next upcoming years.
- 10:38
- Unidentified speaker: And so Austin, I I could go on for other ones, but it's it's really interesting because we we're going to talk about the general state of the market today and what's happening.
- 10:49
- Unidentified speaker: But I think when if you just looked at kind of high-level coverage, what's the headline on CNBC?
- 10:55
- Unidentified speaker: After Alphabet's earnings and after uh Amazon's earnings, you will have said this is a terrible time for AI, and and it was very much framed that way.
- 11:05
- Unidentified speaker: But all of the supplier companies that are now getting kind of taken to the woodshed have have had incredibly bullish earnings and and things can change really fast, and and maybe with some dynamics that are happening in the market.
- 11:17
- Unidentified speaker: The next time they issue quarterly guidance, they'll they'll have taken down some expectations.
- 11:23
- Unidentified speaker: But we see these companies trading for their lowest multiples in a significant amount of time.
- 11:30
- Unidentified speaker: Well, across the board, they're they're talking about um a level of acceleration and expectations for 2027, uh, outlook into 2028 and beyond.
- 11:40
- Unidentified speaker: Um, kind of much the same story we talked about earlier, where there is a lot of pessimism this earnings season, but the actual results continue to be stellar across the board.
- 11:52
- Unidentified speaker: I'm curious, you know, the market is a fickle beast, right?
- 11:56
- Unidentified speaker: And emotions rule as much as rationality, and we know that just from being in this for a long time, as I'm sure our listeners do.
- 12:02
- Unidentified speaker: This feels like, in many ways, I feel like if these same prints occurred last year or the year before, they would have been championed and heralded.
- 12:11
- Unidentified speaker: But now many of these companies are selling off.
- 12:13
- Unidentified speaker: So, what do you think is behind this like discounting of the future?
- 12:17
- Unidentified speaker: Is it now that people and investors were not appreciating how big the investment wave was and now they've caught up?
- 12:24
- Unidentified speaker: So now they want to see current results and the future is being discounted because early in this cycle, you were very pressured and saying, look, you know, all this CapEx money is coming from effectively the hyperscalers and a couple other companies, which will funnel its way into the suppliers.
- 12:36
- Unidentified speaker: Let's position ourselves there.
- 12:38
- Unidentified speaker: That story continues to play out as we've talked about.
- 12:40
- Unidentified speaker: The CapEx numbers continue to go up.
- 12:42
- Unidentified speaker: Meta recently announced they're going to increase CapEx and every other hyperscaler so far continues to be able to push on their inexhaustible levels of cash or financing to do that.
- 12:51
- Unidentified speaker: So that story is intact, which then flows down to the suppliers.
- 12:54
- Unidentified speaker: So, what do you think is behind the sentiment shift here?
- 12:57
- Unidentified speaker: Or is it just that investors are now getting in the show me moment where they want to see current cash flow and current earnings and they are now discounting the future growth that you saw a year ago.
- 13:07
- Unidentified speaker: Do you have a sense?
- 13:08
- Unidentified speaker: Yeah, let's let's walk through that because it's it's really important.
- 13:11
- Unidentified speaker: It's going to kind of determine when we look at a portfolio review and where you'd allocate money today.
- 13:17
- Unidentified speaker: We we really need to walk through this.
- 13:19
- Unidentified speaker: Uh, first of all, yes, so just putting some numbers behind how bad the current momentum sell-off is.
- 13:26
- Unidentified speaker: Uh, the TMT kind of momentum style box, as defined by Goldman Sachs, they have it down 62%.
- 13:35
- Unidentified speaker: And that number is like a week old, so it's it's likely even worse now.
- 13:39
- Unidentified speaker: You know, we could be looking at 65, 66 percent.
- 13:42
- Unidentified speaker: That's actually worse than the dot com crash.
- 13:45
- Unidentified speaker: It's worse than 2022, which was the other recent kind of worst time to be a tech investor.
- 13:51
- Unidentified speaker: It's worse than 2008, and once again, it's worse than the dot-com.
- 13:54
- Unidentified speaker: Um, there's there's some reasons you could call them logical and illogical.
- 14:00
- Unidentified speaker: Uh, you know, uh logical.
- 14:04
- Unidentified speaker: Um, you know, many, there is a fear that we have uh cyclical industries, things like memory reaching their peaks, um, and and people wanting to lighten exposure there, whether or not you believe that's that's right or wrong.
- 14:17
- Unidentified speaker: Um, you you have that, you have capital becoming a key constraint.
- 14:21
- Unidentified speaker: We're gonna dig into that with more detail.
- 14:23
- Unidentified speaker: That creates headwinds and more uncertainty.
- 14:26
- Unidentified speaker: We have the macro situation where not only rising rates leading to less capital available, but also leading to, again, you know, with a company like Amazon or Alphabet, the the biggest issue with their earnings was wasn't the AWS spend or their cloud spend.
- 14:42
- Unidentified speaker: It was the things that are more levered to consumers like search and their e-commerce business.
- 14:48
- Unidentified speaker: We also have herding, where you had retail investors uh really chasing after bottlenecks, uh in many cases rather than more durable trends, and and you have a lot of the air coming out of that idea.
- 15:02
- Unidentified speaker: And then you have less logical ideas.
- 15:05
- Unidentified speaker: Uh, we have record degrossing.
- 15:07
- Unidentified speaker: The amount of leverage that's been removed from the broader market in the past month is once again a historical level.
- 15:14
- Unidentified speaker: We have fund blow ups over the summer.
- 15:17
- Unidentified speaker: One of the key reasons that stocks had tanked so hard, specifically AI stocks, was the amount of capital that um is is often controlled by funds that specialize in AI.
- 15:28
- Unidentified speaker: And and when they blow up, that removes a lot of capital very quickly.
- 15:32
- Unidentified speaker: And and then we have herding, but in the other direction, which is kind of a record flight to safety.
- 15:39
- Unidentified speaker: So just putting some numbers behind it.
- 15:41
- Unidentified speaker: If we if we looked at semiconductors, which has been, to date, the primary way to kind of invest across the AI space.
- 15:50
- Unidentified speaker: It's it's where most of the gains accrued, and um, it's it's where most of the growth was first seen in it.
- 15:56
- Unidentified speaker: So if your glass half empty, semiconductor stocks, they're down 30% from their June high.
- 16:02
- Unidentified speaker: That's gonna be even higher today, maybe more like 35% after the losses we're seeing today, and more than 20% across the past month alone.
- 16:10
- Unidentified speaker: But right, the glass half full is that they're up 50% over the past year and 170% over the past five years, right?
- 16:19
- Unidentified speaker: So you need to weigh both of these issues.
- 16:21
- Unidentified speaker: If if you put all of your chips into the table in April, you have invested in a historically bad market.
- 16:29
- Unidentified speaker: If you did it in April 2025, you're still up 100%, right?
- 16:33
- Unidentified speaker: So so it just that that kind of context matters.
- 16:36
- Unidentified speaker: But the bottom line is what we are seeing is semiconductors being priced more akin to what you normally see at the end of a cycle.
- 16:45
- Unidentified speaker: Um, over the past seven years, they've traded a range of about 12-ish, 12 times to 30 times forward earnings.
- 16:53
- Unidentified speaker: And and right now we're at the lower end there.
- 16:54
- Unidentified speaker: We're at uh, depending on how you measure, maybe 13 times, 14 times forward earnings.
- 17:00
- Unidentified speaker: Um semiconductors have gone from something that were very expensive a year ago to actually cheaper than the broader market.
- 17:10
- Unidentified speaker: Um, and and once again, that's with the expected growth rate next year in terms of earnings for the category of 78%.
- 17:16
- Unidentified speaker: So growing 78%, forward P of uh about 13x, uh, which is which is the lowest levels they've been since 2022.
- 17:27
- Unidentified speaker: So it's it's just interesting because we we have seen this this flight away.
- 17:32
- Unidentified speaker: And as I know, there's a lot of hurting into safety.
- 17:36
- Unidentified speaker: If we look at different categories of the market, consumer staples, 90th percentile, and how expensive stocks are in terms of PE.
- 17:44
- Unidentified speaker: And once again, that that measure is only going to be rising with the kind of action we see today of the Dow outperforming semiconductors by six or seven percent.
- 17:52
- Unidentified speaker: Consumer discretionary, 77th percentile, industrials, 89th, healthcare, 81st percentile, utilities, 84th.
- 17:59
- Unidentified speaker: So Austin, everything conservative, its PE is near the top end of the range that we've seen.
- 18:05
- Unidentified speaker: The only sectors in the market that are below their range are tech, communications, which is mostly tech stocks, it's like Facebook, and then energy.
- 18:16
- Unidentified speaker: So those are the only things that are below their typical PE range.
- 18:21
- Unidentified speaker: So, you know, it's it's one of our listeners, Mike McMahon, he had captured it on um Twitter pretty well, and I I think he was observing this about NVIDIA, but you you look at all this data and you say this is either the fattest pitch that AI bulls have seen within the entire cycle, or the market's gonna look make make me look really stupid a year from now.
- 18:44
- Unidentified speaker: And that that's that's his direct quote there.
- 18:46
- Unidentified speaker: And you know, the the question is what is the market seeing?
- 18:51
- Unidentified speaker: How how can we dig into this more?
- 18:53
- Unidentified speaker: What is this moment that we're at?
- 18:55
- Unidentified speaker: And and I think one of the big things that we need to remember is that people love to compare the AI build-out to the dot-com bust.
- 19:04
- Unidentified speaker: Because I mean, we all know why, right?
- 19:07
- Unidentified speaker: It's just two dominant things that kind of sucked all the oxygen out of the room and and they became the you know, societally dominant trends of their era.
- 19:18
- Unidentified speaker: Uh, but but they are extremely different.
- 19:21
- Unidentified speaker: Because with dot-com the capacity was there, but there was never demand.
- 19:28
- Unidentified speaker: We've talked about the amount of fiber that was built was like two percent of it was used during the dot-com boom, right?
- 19:36
- Unidentified speaker: AI has been fundamentally different.
- 19:38
- Unidentified speaker: Um it's been a continued story of supply racing to catch up with demand.
- 19:44
- Unidentified speaker: Um and and that that's a very different cycle that it is demand driven.
- 19:48
- Unidentified speaker: Uh so at the end of the day, I talked about this in the last episode, but there's there's thousands of data points out there.
- 19:55
- Unidentified speaker: And whether or not AI investing is going to be very profitable over a one-year stretch or a five-year stretch is is really does come down to this demand picture, right?
- 20:08
- Unidentified speaker: And and that's why I talked about when when I start looking at my kind of conviction in AI and the trade itself, I I first start with the capability of AI itself.
- 20:19
- Unidentified speaker: Um, because you know, there's so much noise about picking out individual stocks.
- 20:23
- Unidentified speaker: But the singular question that that is going to answer whether or not we make money on the trade is again, is AI demand going to remain high?
- 20:34
- Unidentified speaker: Um so the near-term risk right now.
- 20:38
- Unidentified speaker: We have we have capital risk.
- 20:40
- Unidentified speaker: Uh that that's the number one uh that's the number one constraint that's happening.
- 20:45
- Unidentified speaker: Austin, you probably saw the quote, but Alon Musk has he he has a way of framing things that are very striking.
- 20:51
- Unidentified speaker: And and this week he predicted guppy level death for for neo clouds.
- 20:56
- Unidentified speaker: I I'm sure you saw that quote.
- 20:58
- Unidentified speaker: Yeah, it was it was pretty striking.
- 21:00
- Unidentified speaker: He uh he's he's he's a perpetual walking headline generator.
- 21:05
- Unidentified speaker: Um you know the this dynamic here reminds me of something you maybe going back to 2014 or 2015.
- 21:11
- Unidentified speaker: Uh I was covering consumer goods stocks at the time, and I remember I don't even know what the crisis du jour was at the time, frankly, um, which probably is is gonna really underscore this point.
- 21:20
- Unidentified speaker: But they started training at a premium.
- 21:22
- Unidentified speaker: I think they, you know, the consumer staple sector went from like 18 times earnings up to like uh 21 and 22 or whatever.
- 21:31
- Unidentified speaker: And you pay a premium for peace of mind, right?
- 21:34
- Unidentified speaker: People want safety when things are uncertain.
- 21:36
- Unidentified speaker: So I watched all this happen, and a lot of investors who that that we were working with at the time who had deeply respect were like, this isn't this isn't gonna end well, and they were right.
- 21:45
- Unidentified speaker: And if you look at the cumulative returns from 2014 to today, the consumer staples sector has underperformed the SP by more than 50%.
- 21:53
- Unidentified speaker: So here's just some quick, some quick numbers just to put that into context.
- 21:57
- Unidentified speaker: Uh, you know, you've got Coke and Pepsi since then, uh, which are sort of classically consumer staples companies up 135%, 105%, Proctor and Gamble up 155%, the XLP, which is the consumer staples sector, spider ETF up 105%, 125%.
- 22:15
- Unidentified speaker: These are decent returns.
- 22:17
- Unidentified speaker: Meanwhile, the SP 500 is up 300% and the NASDAQ up substantially more.
- 22:21
- Unidentified speaker: So you pay a premium for peace of mind, and that premium pays, you know, uh it hurts your portfolio the longer in the future, you know, you you you extend it out.
- 22:32
- Unidentified speaker: And you said that people like to draw conclusions back to the dot-com bubble.
- 22:36
- Unidentified speaker: Sure, it's natural.
- 22:37
- Unidentified speaker: So let's draw another analogy, right?
- 22:39
- Unidentified speaker: We've got the Leopold auction brenner portfolio blow up.
- 22:42
- Unidentified speaker: A lot of people might, you know, if they were bearish on this, would have said, hey, that's the canary in the coal mine.
- 22:46
- Unidentified speaker: Uh the AI trade is about to implode.
- 22:49
- Unidentified speaker: Let's draw an analogy back to 1998, right?
- 22:51
- Unidentified speaker: You've got long-term capital management, one of the largest, most successful hedge funds at the time, also blew up.
- 22:57
- Unidentified speaker: The two years after their blowup were substantially positive.
- 22:59
- Unidentified speaker: I think the market was up 23% and then like 15%.
- 23:02
- Unidentified speaker: Now it eventually did implode, but if you simply held from the point that they blew up and went forward, you still had a 10% IRR every single year, nothing else happening.
- 23:13
- Unidentified speaker: And that beat you know, any consumer staples uh stock.
- 23:16
- Unidentified speaker: So volatility is the price of admission.
- 23:19
- Unidentified speaker: You pay a premium for peace of mind.
- 23:21
- Unidentified speaker: We can draw all of these analogies, and people are so worried about the dot-com blow up.
- 23:24
- Unidentified speaker: But I think like you and I have actually talked about many times.
- 23:27
- Unidentified speaker: There's we remember the bad part of what happened in the dot-com blowup, and very not enough people do the math on what actually happened since.
- 23:35
- Unidentified speaker: I've seen some numbers where if you just like bought at the peak of the dot-com blow-up in the Nasdaq and wrote it down and up, you still beat the SP 500.
- 23:42
- Unidentified speaker: I'm not sure if that's literally true, but you you actually have a meaningfully positive return.
- 23:47
- Unidentified speaker: There's that example with long-term capital management now.
- 23:49
- Unidentified speaker: So, you know, if if you want to draw a parallel to the dot-com blowup and what happened in 99, I get the temptation to do that.
- 23:58
- Unidentified speaker: You need to be honest about, you know, how bad was it really for patient investors?
- 24:02
- Unidentified speaker: And the answer was they still did really, really well.
- 24:04
- Unidentified speaker: And I also don't think that this is that moment.
- 24:06
- Unidentified speaker: That's just painting a worst case scenario analogy.
- 24:10
- Unidentified speaker: Yeah, and when we look about, okay, what are what are the practical implications of kind of this financing shutdown that Musk was warning about?
- 24:19
- Unidentified speaker: Well, it's funny, I I got a few texts from people.
- 24:23
- Unidentified speaker: You had probably seen that open AI, they were changing their fundraising round.
- 24:26
- Unidentified speaker: They were going to previously raise $30 billion.
- 24:29
- Unidentified speaker: They they took it down to targeting something closer to $20 billion with $11 billion secured.
- 24:34
- Unidentified speaker: And it was, oh, okay, this is it.
- 24:37
- Unidentified speaker: This is the end of the AI market.
- 24:39
- Unidentified speaker: They they can't get financing.
- 24:40
- Unidentified speaker: But in the week prior, we had SpaceX close on a $40 billion deal.
- 24:45
- Unidentified speaker: We had a $63 billion syndication for chip purchases for Anthropic, which was backed by Broadcom.
- 24:51
- Unidentified speaker: We had Meta announce plans to raise up to $97.5 billion of equity.
- 24:55
- Unidentified speaker: We had Microsoft announce it would raise $25 billion in uh in in debt, I should say.
- 25:02
- Unidentified speaker: So the the capital is still largely there for the largest companies.
- 25:06
- Unidentified speaker: Now there is an issue that the marginal spend that's going to NeoClouds, that's going to the riskier buyers, like open AI, is closing.
- 25:17
- Unidentified speaker: So that that does carry some risk, even if you do not personally own NeoClouds.
- 25:23
- Unidentified speaker: And the numbers would be that the street, the estimates for hyperscaler CapEx, they're somewhere around 1.2 to 1.3 trillion in 2027.
- 25:32
- Unidentified speaker: You know, that is actually, it has come down, and this is part of what's likely weighing on stocks across the AI spaces.
- 25:38
- Unidentified speaker: We are seeing some estimates come back down, uh especially previous two uh midterms.
- 25:45
- Unidentified speaker: But the NeoClouds themselves, they would contribute about $200 billion on top that $1.2 to $1.3 uh trillion dollar number.
- 25:54
- Unidentified speaker: That gets us to something like $1.5 or $1.6 trillion in spend next month.
- 25:58
- Unidentified speaker: But the the question is if this incremental spend um isn't isn't going to come into the AI space because again, financing is is closed.
- 26:10
- Unidentified speaker: And after midterms, there's idea that building a lot of data science is going to get a lot more challenging.
- 26:15
- Unidentified speaker: Well, this is what's causing a lot of pressure on semiconductor stocks.
- 26:20
- Unidentified speaker: Because again, if if people were buying them for the bottleneck aspect of them, if we have slower growth, once again, that's going to mean that they have less leverage when working with companies like Amazon and with a Microsoft and these large hyperscalers that are buying.
- 26:40
- Unidentified speaker: So this is this is kind of becoming the issue right now, and this is what the market's pricing in beyond things like interest rates, beyond things like macro.
- 26:47
- Unidentified speaker: It is at what point is the slowdown in AI spending enough that it's going to really begin adversely impacting the economics of a lot of these suppliers.
- 26:59
- Unidentified speaker: So that's that's the main question that we have today, right?
- 27:02
- Unidentified speaker: That's that's what the market is ingesting.
- 27:05
- Unidentified speaker: And whether or not the market is going to continue building AI at growth rates, you know, this is a near-term problem.
- 27:13
- Unidentified speaker: But then we have a longer-term issue.
- 27:15
- Unidentified speaker: And this is the other side, too, that the market has largely believed that 2027 is going to form somewhat of a peak.
- 27:23
- Unidentified speaker: And it's not sure if AI spending continues.
- 27:27
- Unidentified speaker: Is that demand still going to be there?
- 27:29
- Unidentified speaker: Because as I talked about in the last episode, coding was the story in 2026.
- 27:33
- Unidentified speaker: Why we saw this ravenous amount of demand for anthropic open AI, we saw the growth rates that they had.
- 27:40
- Unidentified speaker: It's because so much money was pouring into AI use for coding.
- 27:45
- Unidentified speaker: So whether or not you believe that this trend is going to continue into the future is whether or not there's going to be different markets, different use cases that take the baton and keep this next wave of growth.
- 27:57
- Unidentified speaker: I'm old enough to remember when that was the issue in 24, and people were worried if the demand would continue in 25, and then if the demand would continue in 26.
- 28:06
- Unidentified speaker: Now, you know, that's not to say, you know, trees can't grow into the sky.
- 28:08
- Unidentified speaker: It's not to say that there won't be bumps along the way, but this has actually been the biggest concern and criticism of AI from the beginning.
- 28:16
- Unidentified speaker: And yet at every turn, the appetite for this tool and this intelligence has exceeded all bearish expectations.
- 28:21
- Unidentified speaker: This is not the first time the AI industry has been faced with this uh question of whether or not the build-out and the investment can continue and whether or not the demand will be there.
- 28:29
- Unidentified speaker: There are always new applications, and especially for a technology like this.
- 28:34
- Unidentified speaker: And you and I are just now starting to see some of the most promising.
- 28:37
- Unidentified speaker: I don't know.
- 28:37
- Unidentified speaker: We've been sending some information back and forth about life sciences and some of the deep healthcare research that's been going on.
- 28:44
- Unidentified speaker: Citrini had this amazing story.
- 28:45
- Unidentified speaker: I don't know if anybody else saw it where they partnered with Anthropic to uh basically engineer a novel compound on a couple hundred thousand dollars of compute as non, you know, we'll say it like medical researchers, as like, you know, basically amateurs in this industry.
- 29:00
- Unidentified speaker: So um I am not as concerned about this, but what else are you seeing about how this technology will be applied in ways that might increase that, you know, put salve on those concerns about demand perpetuating in 27?
- 29:14
- Unidentified speaker: Yeah, I think one area that there's I would be in disagreement with the broader market is AI's application to knowledge work.
- 29:21
- Unidentified speaker: I think there is a broad consensus that knowledge work for large companies, S P 500 companies, is going to use good enough AI models that won't require cutting edge stuff.
- 29:32
- Unidentified speaker: That's that's going to lead to more demand.
- 29:35
- Unidentified speaker: I think we've seen within our own business that there is always going to be demand for the highest level of intelligence, and there's always going to be a strong ROI on it.
- 29:44
- Unidentified speaker: And when one of the big reasons there hasn't been the broad uptake in knowledge work is once again, agentic AI is less than a year old in terms of its broad applications, and we haven't built kind of the workflows.
- 29:59
- Unidentified speaker: Right now, it takes a lot of creativity from an individual to be able to use AI to solve your use cases.
- 30:06
- Unidentified speaker: I think it's going to become more and more obvious how people are able to use it, both themselves and organizations.
- 30:12
- Unidentified speaker: I think you have other applications too, like research and development.
- 30:16
- Unidentified speaker: You know, we look at obviously there's a lot of money that goes into research and development across the world and how transformative AI can can be in natural world applications.
- 30:30
- Unidentified speaker: Uh it's it's already done that for mathematics, and I think it's going to really expand uh, you know, you think about material sciences and other areas, and and that's an absolutely massive mark.
- 30:39
- Unidentified speaker: And then, like you said, things like life sciences.
- 30:42
- Unidentified speaker: I don't that's one I've struggled with just because it it there's no question that it's going to be extremely capable in that domain, but you have a lot of constraints around human trials and how fast you can move in that industry.
- 30:56
- Unidentified speaker: But again, I think I think that there is more than enough areas for demand to uh continue.
- 31:02
- Unidentified speaker: But Austin, I'd say then, you know, and we can we can talk about stocks.
- 31:06
- Unidentified speaker: I I I wanted to talk about this this broad issue before we start talking about individual stocks, because if you are someone who has AI stocks in your portfolio, the question is we'll we'll get a question that's should should I buy memory stocks today or memory stocks as a group?
- 31:27
- Unidentified speaker: Uh, something that looks attractive today.
- 31:30
- Unidentified speaker: The broader picture is just understanding what the market is doing right now.
- 31:35
- Unidentified speaker: That as far as semiconductors, memory stocks and optic stocks and many of these suppliers are all trading under the same fundamental question, which is how much does demand have to come down in 2027, whether it be through political reasons or whether it be through tightening credit markets, that a lot of these companies begin losing their pricing power, right?
- 32:02
- Unidentified speaker: That that they're at peak margins right now and they're going to lose that ability.
- 32:08
- Unidentified speaker: That's that's more applicable to something like memory, but it also impacts things like optics and and other chip stocks that have been really popular.
- 32:17
- Unidentified speaker: So you want to almost look at your exposure to certain ideas across AI, right?
- 32:23
- Unidentified speaker: And I think this is what we're gonna think about when we're recommending future stocks, right?
- 32:28
- Unidentified speaker: It's it's not that I think maybe an optics company is 15% cheaper than a memory company.
- 32:36
- Unidentified speaker: It's how much exposure I want to this broad concept.
- 32:39
- Unidentified speaker: And do I want new markets, right?
- 32:42
- Unidentified speaker: Would I rather, would I rather look towards getting more exposure to something like robotics that I've gotten a lot more optimistic about the takeoff there?
- 32:50
- Unidentified speaker: Would I would I rather have more exposure to energy?
- 32:53
- Unidentified speaker: Would I rather have more exposure to some software concepts?
- 32:56
- Unidentified speaker: So when I'm looking at portfolio review and when people are looking at their own portfolios, just understanding the underlying drivers that are going to impact broad swaths of your portfolio.
- 33:07
- Unidentified speaker: And if you're too concentrated in one single driver and that's leading to excess kind of volatility, well, then what you're going to want to do is try and diversify out topics, right?
- 33:21
- Unidentified speaker: Okay, well stated.
- 33:22
- Unidentified speaker: And you and you said the key word there, which is why our listeners are here, which is portfolio review.
- 33:27
- Unidentified speaker: So after our you know, sort of industry commentary and earnings coverage, let's actually get into what you recommend people do with their money in their portfolio today, or what are you going to be doing with your portfolio?
- 33:38
- Unidentified speaker: What are some of the big categories and buckets that you can put this money into of where you want to be allocated?
- 33:44
- Unidentified speaker: Then let's look at compute first, right?
- 33:46
- Unidentified speaker: Taiwan Semi, Broadcom, Nvidia, Intel all go in that compute bucket.
- 33:51
- Unidentified speaker: There's others, right?
- 33:52
- Unidentified speaker: There's you know, there's AMD, but they're not in the portfolio.
- 33:55
- Unidentified speaker: Um, there's other chip companies.
- 33:57
- Unidentified speaker: There we've talked about Cerebrus, we've talked about ARM.
- 34:00
- Unidentified speaker: So let's talk about the compute bucket.
- 34:02
- Unidentified speaker: How do you feel about that?
- 34:04
- Unidentified speaker: Well, I think within the compute bucket and the stocks that we've recommended, you're gonna want to look at a continuum of risk and reward at the current point.
- 34:12
- Unidentified speaker: Um is buying Taiwan semiconductor maybe the most opportunistic company if we see a bounce back right now?
- 34:18
- Unidentified speaker: No, it's not.
- 34:20
- Unidentified speaker: But also, this is a company that I don't think there's ever really a bad time to add Taiwan semiconductor.
- 34:26
- Unidentified speaker: The the expectation is they'll do more than $150 billion in net profits by 2028.
- 34:32
- Unidentified speaker: Uh, even with some of the headwinds maybe to DAS AR construction, I think, I think that's a number they're going to be around, and that's against a $1.8 trillion market cap.
- 34:42
- Unidentified speaker: So uh again, this is if you've got consumer staple stocks trading at 17, 18 times their 2028 levels, or you have the most dominant technology brand in the world trading increasingly close to 10 to 11 times earnings for the exact same time.
- 35:00
- Unidentified speaker: Um, that's that's definitely something that I would want to have in my portfolio.
- 35:04
- Unidentified speaker: So I I feel very good about a Taiwan semiconductor.
- 35:08
- Unidentified speaker: Uh they they feel like they're at attractive valuation relative to where they've been.
- 35:13
- Unidentified speaker: And if it's not something that someone has in their portfolio, I would definitely begin there just because it can be such an anchor stock.
- 35:19
- Unidentified speaker: Uh NVIDIA, I I could see the possibility to take some downward revisions in the next year because again, this is a company that has a little bit more exposure to neo clouds.
- 35:32
- Unidentified speaker: If financing is really getting tight, um, you know, this is this is going to impact something like NVIDIA more.
- 35:38
- Unidentified speaker: But again, expectations for next year are $16 in EPS.
- 35:42
- Unidentified speaker: Uh if if they lost 7 or 8% of their total demand due to neo cloud financing, they're they're still trading at 13 or 14 times forward earnings.
- 35:54
- Unidentified speaker: Uh that doesn't mean that there won't be any near-term headwinds, but on a long-term time frame, it feels like a very good time still to buy an NVIDIA.
- 36:04
- Unidentified speaker: I would say if we looked at a Broadcom, they're going to be more riskier than NVIDIA because they have more exposure to financing with their plans about their very large projects with companies like Anthropic.
- 36:17
- Unidentified speaker: But again, trading in a similar range and the highest risk reward of the group, Austin, and it's one that you said you had recently bought uh that just gets more attractive as it falls, is Intel.
- 36:28
- Unidentified speaker: As we had noted last week, they clearly had a misstep in terms of how they allocate between the DAS AR market and PC market, but nothing changes that for this stock to be an outsize winner, it has to get foundries right.
- 36:44
- Unidentified speaker: They said absolutely nothing about foundries and they lost 60% of their value.
- 36:49
- Unidentified speaker: So if if if Intel is a stock that you haven't been invested in and you you want to take some kind of position in it, uh once again, I think it's getting much more attractive, and you need to be aware of the risk that it's facing.
- 37:04
- Unidentified speaker: But again, this is this is a five-year play in terms of the fact that the United States is is going to have every incentive in the world to build a domestic foundry.
- 37:13
- Unidentified speaker: And and very little's changed on that unless um you you have a lot less confidence in this management team.
- 37:19
- Unidentified speaker: So out of that group, I would say if if you want the highest risk reward right now, that's that's Intel because they've been sold off 60%.
- 37:27
- Unidentified speaker: Well, their long-term picture uh has changed very minimally in many ways.
- 37:33
- Unidentified speaker: And if you want the the stock that's probably the safest one trading at a relatively uh attractive valuation right now, I would look to Taiwan semiconductor.
- 37:43
- Unidentified speaker: Yeah, you know, speaking of Intel, it is it's it's their game to win here.
- 37:47
- Unidentified speaker: I mean, as you have you discussed on many episodes in our episode last week as well, they are chronically execution challenged, but they have so much upside and potential you sort of need to bet there.
- 37:56
- Unidentified speaker: That was the same story with Marvell.
- 37:58
- Unidentified speaker: They were a dog for the first year or so in the portfolio before finally taking off.
- 38:02
- Unidentified speaker: Uh, they're leaning into the interconnect business.
- 38:04
- Unidentified speaker: Uh, I saw with the Lumentum acquisition.
- 38:07
- Unidentified speaker: Intel feels like it's similar.
- 38:08
- Unidentified speaker: There's so many things that they can do as long as they execute.
- 38:12
- Unidentified speaker: Uh, long-term, I don't know how we don't see a separation of the foundry business and the chip business, but you know, who who knows how long that could take.
- 38:19
- Unidentified speaker: There's a lot of feelings tied up there.
- 38:21
- Unidentified speaker: So uh I like Intel for the risk reward profile, but just full disclosure, it's going to be a white knuckle roller coaster ride, and you should expect that.
- 38:28
- Unidentified speaker: So I did buy more.
- 38:29
- Unidentified speaker: It's not a huge position in my portfolio, but if you want exposure to this space, I do like it.
- 38:33
- Unidentified speaker: And I think if they can get the execution uh lined up in the same way Marvell did, it can do quite well.
- 38:39
- Unidentified speaker: Uh let's look at memory, one of the biggest trades in the AI industry, uh, right after optics.
- 38:45
- Unidentified speaker: So we've got SK Heinex, Anto, ACMR, Micron.
- 38:50
- Unidentified speaker: Um, how do you feel about this industry now?
- 38:52
- Unidentified speaker: I think I saw Micron, you know, for all the sell-off and weakness, they were sort of weak early on.
- 38:57
- Unidentified speaker: And I think they're now maybe like they've been hanging around like 950 a share for a couple months now, while a lot of other stuff has been selling off.
- 39:02
- Unidentified speaker: I'm not sure if they're still there.
- 39:04
- Unidentified speaker: Um, but they have they have felt a little more.
- 39:07
- Unidentified speaker: Oh, I lied.
- 39:08
- Unidentified speaker: They're at 800 a share.
- 39:09
- Unidentified speaker: Maybe they did sell off.
- 39:10
- Unidentified speaker: Sorry.
- 39:10
- Unidentified speaker: Stuff has moved quite a bit.
- 39:12
- Unidentified speaker: Looks like I missed the last month there.
- 39:14
- Unidentified speaker: Um, okay.
- 39:15
- Unidentified speaker: Well, then let's talk about the memory trade.
- 39:17
- Unidentified speaker: How do you feel about this industry right now?
- 39:20
- Unidentified speaker: Is this like Intel where maybe people are getting a volatile sector at a discounted price?
- 39:26
- Unidentified speaker: Or were the naysayers right and cyclicality is finally going to catch up with all of the optimists and wipe them out?
- 39:34
- Unidentified speaker: Just being like 100% full disclosure here, this is this is it's it's funny because when we recommend it, it's always interesting because you're talking about whether or not people want to own more of something or buy more of something in this point in time.
- 39:47
- Unidentified speaker: And and then we're also talking about where it stood when we first recommended it.
- 39:51
- Unidentified speaker: So for all these companies, it was a completely different world when we first recommended them, right?
- 39:55
- Unidentified speaker: So uh a micron was $80 a share and and it's $800 a share today or whatever.
- 40:01
- Unidentified speaker: I uh I would put memory and optics in an extremely close bucket to each other right now.
- 40:07
- Unidentified speaker: They they have some different, they they have some differences to their profiles.
- 40:11
- Unidentified speaker: Uh, you know, memory is seen as something how long can it hold on to its peak, whereas optics is seen as something that's going to be able to see substantial growth in the years ahead as it becomes a key differentiator in AI networking.
- 40:25
- Unidentified speaker: But these are two industries that when you go back through time are extremely cyclical, right?
- 40:32
- Unidentified speaker: It's always been memory goes through waves and the fear is when this wave is going to end.
- 40:38
- Unidentified speaker: Optics is very similar.
- 40:39
- Unidentified speaker: If you go back and you look at the share charts for a company like Coherent, um, when when when oversupply situations come, they can be incredibly brutal to both of these industries.
- 40:50
- Unidentified speaker: So they're both trading on a similar dynamic that the market is essentially trying to arrive at some kind of number.
- 40:58
- Unidentified speaker: What kind of growth rate are we going to see in 2027 and how is this going to impact that supply situation?
- 41:06
- Unidentified speaker: Um, I I think just more research is going to be needed on it.
- 41:09
- Unidentified speaker: But I I will say I had said personally I was lightening up my optics exposure back in June because I thought that the stocks they had just run up quite a bit.
- 41:20
- Unidentified speaker: And I I just had too much optics in my portfolio, right?
- 41:24
- Unidentified speaker: And for the AI investor portfolio, the question is do we have too much relative exposure to these kind of cyclical uh semiconductor spaces?
- 41:35
- Unidentified speaker: We we they've been massive winners since we first recommended them.
- 41:38
- Unidentified speaker: But as we move into the future phases of AI, do we want to keep exposure at the current level that we have, or do we want to reallocate to other areas?
- 41:47
- Unidentified speaker: I haven't quite made a decision on that because again, there's we we we just had midterms yesterday, which is going to be an incredibly um pivotal area.
- 42:00
- Unidentified speaker: I mean, it could it could be nothing and and it could be significant in terms of of how much this is going to affect the build-out.
- 42:06
- Unidentified speaker: And and there is a very complex analysis to see what different levels of spend will be next year relative to the supply that's being brought on across these categories.
- 42:15
- Unidentified speaker: But I will say the Marvel, uh Marvell, I should say, Lumentum buyout.
- 42:21
- Unidentified speaker: I I saw a lot of people reading that as a lumentum cashing out.
- 42:25
- Unidentified speaker: Well, I I don't think that quite captures the situation.
- 42:29
- Unidentified speaker: I think uh if they were cashing out, they wouldn't have done an all-stock deal.
- 42:32
- Unidentified speaker: They wouldn't have joined the board for Marvell.
- 42:34
- Unidentified speaker: Um, I think this is an attempt to get an incredible amount of vertical integration to compete against a company like Broadcom in the space.
- 42:43
- Unidentified speaker: And I think this is a real all-in on the fact that Marvell's management team believes interconnects are going to be one of the key growth stories of the next five years.
- 42:52
- Unidentified speaker: So it seems like a fantastic deal for Marvell, a less fantastic deal for Lumentum.
- 42:58
- Unidentified speaker: But if you own Lumentum, you're about to have more Marvell shares because it is an all-stock deal.
- 43:03
- Unidentified speaker: Um in terms of other companies from the space that I do think are attractive relative to recent sell-offs.
- 43:11
- Unidentified speaker: I recommended Semtech multiple times.
- 43:13
- Unidentified speaker: It is on the more expensive end, but I really like their dynamics, their competitive position, how they work in both um copper and the optics industries, how they get a higher, how they get a higher revenue base as we continue moving faster and faster to more advanced networking speeds.
- 43:32
- Unidentified speaker: And and I like their exposure to TPUs.
- 43:34
- Unidentified speaker: Uh I I I mentioned earlier, we we're probably going to have a situation where if there is a slowdown in spending that is in ways caused by government constraints, uh, well, that that that that's going to impact a company like Google probably less.
- 43:54
- Unidentified speaker: They're going to be incredibly capable about being able to work with local governments on projects.
- 43:59
- Unidentified speaker: It means that they're going to continue building because they're not going to have to go so far into debt financing.
- 44:04
- Unidentified speaker: And they're going to have a runway to grow their TPU franchise for years ahead.
- 44:08
- Unidentified speaker: One more stock I want to talk about, ACM research.
- 44:11
- Unidentified speaker: It is down 50% in the past month.
- 44:15
- Unidentified speaker: Just uh a parabolic move downwards.
- 44:17
- Unidentified speaker: I need to do some more research on that, but they have a lot of exposure to China.
- 44:23
- Unidentified speaker: The sentiment for AI in China is if you think it's bad in the United States right now, it is incredibly worse there.
- 44:30
- Unidentified speaker: So this could be something that it is a temporary opportunity.
- 44:35
- Unidentified speaker: It could be it could be something that uh the situation over in China has materially gone worse.
- 44:40
- Unidentified speaker: I haven't fully ingested that, but I'm going to do some more research on that.
- 44:44
- Unidentified speaker: And then I think the last category that I'm just really interested across here is continuing to be advanced packaging.
- 44:50
- Unidentified speaker: I talked earlier about error test systems and the validation from new customers uh beginning to do large projects uh with their test equipment, but we also have companies like Anto and Camtech.
- 45:02
- Unidentified speaker: And again, in general, you I really would like to own more stocks where the the the wind is at their back.
- 45:11
- Unidentified speaker: And and we just see the trends for advanced packaging to become more and more material as a percent of semiconductor spend.
- 45:19
- Unidentified speaker: So if if we did see lower growth rates, I and I'm getting much more attractive valuations on the companies that are in the industries that can grow significantly faster than broad AI, I'm gonna use an opportunity like the current sell-off to kind of explore whether or not we want to continue adding there.
- 45:38
- Unidentified speaker: And and if we add, whether or not we do that by selling other stocks or we're just going to establish new positions.
- 45:45
- Unidentified speaker: Okay, great analysis there.
- 45:47
- Unidentified speaker: Let's move on to a sector we have talked about a lot, but that doesn't have an enormous position in the portfolio.
- 45:54
- Unidentified speaker: It has some exposure, you're building it out, but I want to talk about robotics.
- 45:57
- Unidentified speaker: You know, it's my favorite.
- 45:59
- Unidentified speaker: My view of this has been this industry will be more consequential than people expect, but it will take longer than early pricing suggested.
- 46:07
- Unidentified speaker: Uh valuations overall are pretty high in this space, or at least they were, and they're looking a lot better now.
- 46:13
- Unidentified speaker: So you've got Allegro Microsystems, Rex or Regal Rex Nord, Ouster, which we talked about earlier.
- 46:19
- Unidentified speaker: How is robotics looking today?
- 46:21
- Unidentified speaker: Are we getting more favorable valuations with this sell-off?
- 46:24
- Unidentified speaker: Are you thinking about expanding the robotics sleeve, or is this still a wait and see because the runway is going to be so long here?
- 46:32
- Unidentified speaker: Yeah, as we've talked about robotics, it's always a question of how quickly is this coming?
- 46:37
- Unidentified speaker: Um, you don't want to invest too early, but if you if you continue to wait until it's it's here and it's obvious, you're you're going to miss it.
- 46:46
- Unidentified speaker: And the biggest thing that we wanted was better valuations on many of these companies, and we are getting better valuations on them.
- 46:54
- Unidentified speaker: So when I look at where I want to deploy capital today, if the question is, are you too indexed to many of the supplier stocks relative to other trends and and concepts within AI, a good place to look today is robotics, especially if we are getting many of these companies at kind of the cheapest levels we've seen since the beginning of 2026.
- 47:17
- Unidentified speaker: So I I think what I've seen robotics, we did the deep dive on it.
- 47:22
- Unidentified speaker: It's it's incredibly promising.
- 47:24
- Unidentified speaker: It's always been a very fragmented space, but we're seeing a lot more generalization from foundation models.
- 47:30
- Unidentified speaker: Uh data scaling is very early, but promising.
- 47:33
- Unidentified speaker: Also, another thing that we're seeing, the frontier models, such as GPT-6 Astra, they're they're becoming very capable at robotics, despite not being robotics foundation models, or or even we don't know the level of training that they've had on any robotics data.
- 47:48
- Unidentified speaker: We've seen similar advancements from uh Google's Argon.
- 47:51
- Unidentified speaker: So I I think we're now seeing multiple scaling vectors across the space.
- 47:58
- Unidentified speaker: And what we're gonna have an opportunity to do for robotics is we're going to be able to look at, you know, when I recommend Regal Rex Nord, I said, well, this isn't this isn't the most high upside concept, but they have a very stable business at their foundation.
- 48:17
- Unidentified speaker: So, you know, they're they're not going to have the same downside as some semiconductor company that that's relying on 100% growth and and and if it sees a big deceleration could fall 60, 70 percent.
- 48:30
- Unidentified speaker: It it it dropped after we recommended it because everything across the AI space has has had terrible sentiment.
- 48:38
- Unidentified speaker: But I do think there's more of a limited downside in a concept like that.
- 48:42
- Unidentified speaker: So we're gonna be able to look at some of the kind of industrials adjacent companies.
- 48:47
- Unidentified speaker: We're gonna look at some companies that have more specialization in robotics.
- 48:51
- Unidentified speaker: And we could also look at some companies that fit a profile more akin to ouster.
- 48:57
- Unidentified speaker: You know, a lot of a lot of the more high growth concepts in robotics are 60, 70, sometimes 80% cheaper than they were.
- 49:05
- Unidentified speaker: So we don't want to introduce undue risk, but it's also the most attractive time to look at them in a long time, as in many ways, robotics is at its most promising point.
- 49:15
- Unidentified speaker: So I think if if we're looking at what concepts we're going to add coming up as we rebalance the portfolio and think about how we want the portfolio to reflect what the next one or two years of AI looks like, I think the amount of robotics we have will at a minimum double.
- 49:32
- Unidentified speaker: Um so I think this is going to be a key focus area.
- 49:36
- Unidentified speaker: I love to hear that.
- 49:37
- Unidentified speaker: I'm very excited for those picks.
- 49:39
- Unidentified speaker: Um, I want to look at one of the smallest sleeves of the portfolio, one that I have the least understanding of and I'm very confused about.
- 49:45
- Unidentified speaker: So we've got the skilled labor companies, which is IES Holdings and Mcore, tickers IESC and EME.
- 49:52
- Unidentified speaker: I don't understand what's happening.
- 49:54
- Unidentified speaker: IES Holdings is down 17% today, MCore is down 11.
- 49:58
- Unidentified speaker: Now we are filming on Wednesday the fourth.
- 50:00
- Unidentified speaker: This looks to be some post-election reaction, I think.
- 50:03
- Unidentified speaker: But can you explain to me why these companies are down so much?
- 50:07
- Unidentified speaker: Because I'm I'm not following the pattern here.
- 50:09
- Unidentified speaker: Yeah, this is this is definitely a reaction.
- 50:12
- Unidentified speaker: Let's let me look at how much did you say they were down today?
- 50:15
- Unidentified speaker: Uh MCore is down 11 today, and IE, well, as of the time of the filming, and IES holdings is down 16 and a half.
- 50:24
- Unidentified speaker: Clearly looks to be some post-election overhang.
- 50:27
- Unidentified speaker: But what's surprising is that both of these companies felt like the more sort of stable ones where you were paying a little bit of a premium, getting a higher multiple, but they were going to be less volatile than some of the other names in your portfolio.
- 50:38
- Unidentified speaker: And yet, and yet they're down this much.
- 50:40
- Unidentified speaker: I'm not sure I follow the king the reasoning here.
- 50:44
- Unidentified speaker: So can you explain to me one, how do you feel about skilled labor companies in the portfolio overall?
- 50:49
- Unidentified speaker: And then just what's going on today?
- 50:51
- Unidentified speaker: Yeah, as you mentioned, the the hope for for them, we had recommended them when we said we thought the market was a little bit too hot.
- 50:58
- Unidentified speaker: I think we'd recommend them in May, and we were concerned about where valuations were reaching.
- 51:02
- Unidentified speaker: Our hope was they would have less volatility than some other options while continuing to give upside to the AI build-out.
- 51:09
- Unidentified speaker: These are companies that have just generally, as a group, performed excellently across the past 10 plus years.
- 51:16
- Unidentified speaker: Uh it's just been a very solid industry.
- 51:18
- Unidentified speaker: They've been following a game plan of continuing to consolidate.
- 51:22
- Unidentified speaker: And I'm a little confused by this pricing today as well, because as you mentioned, they this is a midterm reaction.
- 51:29
- Unidentified speaker: The concept here is that there's going to be less data center starts uh with again, you know, a Democrat winning in uh Texas and and also many of the kind of ballot initiatives that have passed.
- 51:42
- Unidentified speaker: But but these companies, you you look at them, you look at the pricing dynamics in memory, right?
- 51:50
- Unidentified speaker: There have been past memory cycles where memory went from 60% gross margins down to negative gross margins.
- 51:58
- Unidentified speaker: And and these names are trading in a far more exaggerated way than something like memory.
- 52:04
- Unidentified speaker: You look at labor supply, it's a whole heck of a lot harder to get incremental labor supply than build a new memory factory, right?
- 52:13
- Unidentified speaker: You can you can increase the supply of a certain chip line 50% a year.
- 52:18
- Unidentified speaker: The amount of labor supply that we have for building out data centers has been chronically undersupplied.
- 52:25
- Unidentified speaker: And if our growth rate next year goes from uh 60% down to 40%, I don't think this crashes their ability to get favorable, their ability to get favorable pricing across these projects.
- 52:38
- Unidentified speaker: So this looks like an extreme overreaction to me.
- 52:40
- Unidentified speaker: I think Comfort Systems was also down more than 15%.
- 52:44
- Unidentified speaker: Um, so you it's always hard to go and buy a company after it drops 50%.
- 52:50
- Unidentified speaker: You're gonna need some confidence.
- 52:52
- Unidentified speaker: Is there something that I'm missing here?
- 52:54
- Unidentified speaker: But when I look at these stocks trading with similar dynamics to many other stocks in the portfolio that that have some of the most cyclical qualities to them, and it seems like a whiff from the market to me.
- 53:07
- Unidentified speaker: And and that makes me more interested in potentially adding to this category.
- 53:11
- Unidentified speaker: Yeah, definitely feels like a misread, uh, some sort of knee-jerk reaction, which will probably be hopefully unwound in the next few months because it just doesn't match my expectation of what these companies are and what they do and where they'd be positioned in the marketplace.
- 53:24
- Unidentified speaker: Also, like just at a minimum, uh data, more data centers, less data centers, uh, wherever the winds blow the next few years, these are like there's universal consensus that like the grids need to be upgraded and electricity demand is going to go up.
- 53:38
- Unidentified speaker: Like that there's you know, there's I don't know.
- 53:40
- Unidentified speaker: I I don't get it.
- 53:42
- Unidentified speaker: But let's move on to semiconductor equipment.
- 53:45
- Unidentified speaker: Talk to me about if if we're seeing a lot of concern about chip manufacturing, presumably semiconductor equipment should follow suit.
- 53:54
- Unidentified speaker: Again, not saying we support that view, but that's some of the views that you know people are holding in the market.
- 53:58
- Unidentified speaker: So, how do you feel about that sleeve in the portfolio right now?
- 54:02
- Unidentified speaker: I you know, it's it's one of those things.
- 54:04
- Unidentified speaker: Again, we're we're having to ingest it a podcast, it's such a poor format for this kind of discussion because it's like these deeply technical, it's almost like a spreadsheet discussion, looking at where growth rates are going to come in in the next few years.
- 54:18
- Unidentified speaker: It's funny, I've talked a lot on this podcast that I I hated seeing CapEx go from eight, you know, 800 billion last year up to 1.5 trillion because I was like, it's too much growth too fast, right?
- 54:33
- Unidentified speaker: It's not just the bubble aspect.
- 54:35
- Unidentified speaker: I've brought up the comparison before of Apple and the level of smartphone growth versus Google, which had just kind of a consistent growth rate through search and and how that made the ride up so much less volatile.
- 54:47
- Unidentified speaker: They both got to approximately the same returns, but owning Apple mid-last decade was gut-wrenching at times.
- 54:55
- Unidentified speaker: I I trust a number of people listening to this podcast had owned Apple.
- 55:00
- Unidentified speaker: And what we're seeing right now is that extreme growth rate for semiconductors, it is going causing this gut-wrenching thing.
- 55:09
- Unidentified speaker: Is now the question is, well, are we actually not going to see the growth rate that was planned, right?
- 55:15
- Unidentified speaker: And at the end of the day, it will likely be better for this industry across a five-year time frame in many ways, right?
- 55:24
- Unidentified speaker: If if if we have a persistent situation of where demand continues to exceed supply for serving AI, that's going to be better across five years.
- 55:35
- Unidentified speaker: But in the near term, there are some just brutal reactions to it.
- 55:39
- Unidentified speaker: One of them is in semiconductor equipment.
- 55:41
- Unidentified speaker: Uh, Austin, I want to go back and see where this stock traded one second when we talked about ultra-clean holdings.
- 55:48
- Unidentified speaker: This is one that we have had listeners ask us about.
- 55:51
- Unidentified speaker: It's a semiconductor equipment player.
- 55:53
- Unidentified speaker: They make subsystems that that go to much larger customers.
- 55:57
- Unidentified speaker: We had been asked about, I said, I like it.
- 56:00
- Unidentified speaker: It's an extremely solid business.
- 56:01
- Unidentified speaker: It's just too expensive.
- 56:03
- Unidentified speaker: Well, I'm looking at right now, it's down 70% from where traded in June.
- 56:08
- Unidentified speaker: Uh, it traded for more than $150.
- 56:11
- Unidentified speaker: Well, $144 per share.
- 56:13
- Unidentified speaker: Now it's down to $14 or $44 per share.
- 56:17
- Unidentified speaker: What we're looking at is a stock that's now trading for about 7x its 2027 earnings.
- 56:23
- Unidentified speaker: And with semiconductor equipment, what you never want, again, is this this crash.
- 56:30
- Unidentified speaker: Semiconductor equipment's been an extremely fabulous category.
- 56:35
- Unidentified speaker: We've talked about before on the show.
- 56:36
- Unidentified speaker: If you go back the past 15 years, three of the top performing stocks were in semiconductor equipment, yet no one could name them, right?
- 56:44
- Unidentified speaker: They are the furthest thing from household names.
- 56:47
- Unidentified speaker: And what semiconductor equipment would want along a five-year time frame is a more consistent climb.
- 56:53
- Unidentified speaker: They they do not want these excesses.
- 56:56
- Unidentified speaker: They do not want a situation where they are having way too much supply relative to the industry.
- 57:03
- Unidentified speaker: They would like this climb.
- 57:04
- Unidentified speaker: So I think for a company like Ultra Clean Holdings, I mean, it it might be that we see how this situation is developing is actually better for the stock across a multi-year timeframe.
- 57:18
- Unidentified speaker: But in the current environment, it is getting sold off radically, down to, again, seven times 2027 earnings.
- 57:25
- Unidentified speaker: And maybe those earnings estimates come down a little bit, but it's it's still extremely cheap, multiple.
- 57:30
- Unidentified speaker: So I'm just weighing right now.
- 57:32
- Unidentified speaker: I was going to recommend this stock today.
- 57:34
- Unidentified speaker: I was going to add it, and and and I decided I'd hold off.
- 57:38
- Unidentified speaker: We'll do a little bit more of a we'll do a little bit more of a new recommendations episode next week.
- 57:44
- Unidentified speaker: But just with what's going on with the market today and how volatile it was, I want to take a little bit more time to think about how we wanted to package up some different recommendations.
- 57:52
- Unidentified speaker: But again, it it just illustrates how cheap some of these concepts have gotten that we were talking about it, we were talking about it in June.
- 58:02
- Unidentified speaker: And again, I said it was an excellent company, it was just too expensive.
- 58:06
- Unidentified speaker: What is the downside that we're seeing today?
- 58:08
- Unidentified speaker: Is is the downside, we just need to bring it back to that original question I was asked about.
- 58:14
- Unidentified speaker: Is the downside that AI itself is fundamentally broken?
- 58:18
- Unidentified speaker: Is the downside that we are not going to see demand for AI?
- 58:22
- Unidentified speaker: Um, as of right now, those things seem relatively in place.
- 58:26
- Unidentified speaker: The biggest issue is just whether or not we are going to see a lower growth rate next year, largely from some external factors.
- 58:34
- Unidentified speaker: Number one, financing.
- 58:35
- Unidentified speaker: Number two, the political situation, which is increasing this week.
- 58:39
- Unidentified speaker: Nothing's changed for the core concept, though, that if AI, if the actual systems behind AI, the foundation models continue improving at their current rate, we are going to have a situation where there's more demand than supply.
- 58:54
- Unidentified speaker: So whether that not that's a 60% growth rate versus a 50% in 2027, leading to maybe a 25% growth rate in 2028 instead of a 20%.
- 59:06
- Unidentified speaker: These are moving around, but the long-term picture isn't necessarily changing a whole lot here.
- 59:11
- Unidentified speaker: So if you're able to get leading stocks across the space for a 70% discount, well, I I am going to want to make that move, right?
- 59:21
- Unidentified speaker: And and maybe, maybe we will see.
- 59:24
- Unidentified speaker: We we have gotten a little overly concentrated in some of the cyclicals, um, and and we want to reduce our exposure there just for the sake of building a more resilient portfolio without volatility.
- 59:36
- Unidentified speaker: But it it just some of these semiconductor companies, semiconductor equipment companies, they they show just how much the markets change for companies that I do believe are still fundamentally about as good as they are today when they're trading for uh 70% more back in June.
- 59:54
- Unidentified speaker: Yeah, there is absolute, there are absolutely some fantastic opportunities in the market.
- 59:59
- Unidentified speaker: As I mentioned in the last episode, and I think you and I can probably both speak to.
- 1:00:02
- Unidentified speaker: Um, we've been doing some shopping the last week, and I'm sure there will be more positions added or you know adjusted in the AI portfolio based on some of these categories that you've talked about.
- 1:00:11
- Unidentified speaker: So the opportunities today on a valuation basis are much better than what we saw six months ago, but in many cases, the trends behind them are still holding, right?
- 1:00:20
- Unidentified speaker: That this the same dynamic held that pushed them to those initial valuations.
- 1:00:24
- Unidentified speaker: So although I don't love seeing my portfolio in the red, I do love the opportunities that we're being presented with in some of these spaces today.
- 1:00:31
- Unidentified speaker: Um, Eric, any parting thoughts for our listeners here?
- 1:00:33
- Unidentified speaker: You talked about a lot of categories.
- 1:00:34
- Unidentified speaker: Thank you.
- 1:00:35
- Unidentified speaker: I know this is not easy.
- 1:00:36
- Unidentified speaker: You know, to our listeners, I know Eric has to spend an incredible amount of time going through each of these sectors to understand the tidal patterns and where trends are moving.
- 1:00:43
- Unidentified speaker: So you threw a lot out there.
- 1:00:44
- Unidentified speaker: You've got compute, memory, optics, robotics, skilled labor, semiconductor.
- 1:00:49
- Unidentified speaker: We touched on other things as well.
- 1:00:50
- Unidentified speaker: Is there any other like parting thoughts you want our listeners to hear today as they're you know looking at their portfolio, experiencing more uh losses than they probably have over the last few years?
- 1:01:01
- Unidentified speaker: I won't say volatility because volatility is both directions, but maybe experiencing some unusual losses.
- 1:01:06
- Unidentified speaker: What would you say to our listeners?
- 1:01:09
- Unidentified speaker: Um, you know, I I would just say it it it is it is always very hard.
- 1:01:14
- Unidentified speaker: Uh I there's been times where the market dropped and we came on and I was just like, I am so excited right now.
- 1:01:22
- Unidentified speaker: I am so excited.
- 1:01:23
- Unidentified speaker: This this one is this is very brutal.
- 1:01:26
- Unidentified speaker: This has been since we started the show.
- 1:01:28
- Unidentified speaker: It's it's up there with Deep Seek was a real kind of gut check moment.
- 1:01:33
- Unidentified speaker: Um, the tariff tantrum was kind of uh a gut check moment, and it forces you to go back and revisit your priors and and say, are you missing something?
- 1:01:43
- Unidentified speaker: I I think I I probably when I filmed the last episode, I was probably a little too blasé, or I didn't touch on enough some of the downstream implications of financing, right?
- 1:01:54
- Unidentified speaker: It it's it's easy to say, well, that's mostly going to hit this group of neo-cloud companies.
- 1:02:01
- Unidentified speaker: It's it's harder to work out how it flows through the entire system and whether or not it's right or wrong, why the market is looking at that and ingesting how it's valuing other companies that are a couple steps away.
- 1:02:14
- Unidentified speaker: And and then across the past week, I've I've been able to put more of those connections together.
- 1:02:19
- Unidentified speaker: But in an extremely dynamic markets, the other side is it's it's it's when it means the most to sharpen your pencil and go through and look at everything.
- 1:02:29
- Unidentified speaker: So, you know, I probably spent 20 hours going through and looking through the portfolio and trying to understand why stocks are moving the way that they are and it what I'm potentially missing.
- 1:02:40
- Unidentified speaker: And and I think the next step again is is trying to make sure that the portfolio is going to properly reflect where we think the opportunities are across the next one to two years.
- 1:02:50
- Unidentified speaker: And if that involves some selling, you know, we'll be ready to do that.
- 1:02:55
- Unidentified speaker: If if it involves buying, we're we're certainly going to buy some more stuff and and try and be opportunistic about this moment.
- 1:03:02
- Unidentified speaker: But you know, the bottom line, Austin, as I've said before, spending all this time looking at AI, I really do think it comes down to the one singular question.
- 1:03:12
- Unidentified speaker: That is, is demand going to stay elevated in 2027, 2028?
- 1:03:19
- Unidentified speaker: And it that will flow from the capability of AI.
- 1:03:22
- Unidentified speaker: If that happens, I I think staying the course is going to be richly rewarded.
- 1:03:28
- Unidentified speaker: And and from everything that I've seen, I haven't seen command in the past three months as you know, we've really seen stuff go haywire.
- 1:03:37
- Unidentified speaker: I haven't seen that core issue of the models and the capability of AI go off any kind of trend line that was on previously.
- 1:03:46
- Unidentified speaker: So, you know, that that heartens me.
- 1:03:48
- Unidentified speaker: And uh we we will come back next week.
- 1:03:51
- Unidentified speaker: We will have some of those additional recommendations.
- 1:03:53
- Unidentified speaker: But I just want to also thank everyone.
- 1:03:55
- Unidentified speaker: You know, if you are watching today, we understand how hard this is.
- 1:03:59
- Unidentified speaker: This is not fun, right?
- 1:04:00
- Unidentified speaker: Uh the amount of text I get from people about AI has probably down about 90% right now.
- 1:04:07
- Unidentified speaker: There's there's just very few people interested in it.
- 1:04:10
- Unidentified speaker: So staying engaged right now, it's not easy, but we we appreciate everyone being on this journey because uh as we've said from the beginning, we believe 10 years from now you'll look back and say, why wasn't it obvious that I did that?
- 1:04:25
- Unidentified speaker: Of course, AI was going to be the next big thing.
- 1:04:27
- Unidentified speaker: Of course, it created these $20 trillion companies and and had this impact on the economy.
- 1:04:33
- Unidentified speaker: But doing it in the moment and going through this process is just two different things.
- 1:04:38
- Unidentified speaker: And and that's why we create the podcast, because even even the greatest investments of all time, an Amazon fell 90%, and it fell 60%, it fell 40%.
- 1:04:50
- Unidentified speaker: It's NVIDIA has been a complete roller coaster at times.
- 1:04:54
- Unidentified speaker: So it's never as easy as it looks, it's never as easy as the 30,000% return you see on a uh piece of marketing copy.
- 1:05:02
- Unidentified speaker: Uh it's very hard.
- 1:05:04
- Unidentified speaker: And um we appreciate everyone.
- 1:05:06
- Unidentified speaker: And also, we would if anyone has any comments, if if they want us to cover anything, I didn't see as many last week, which I was kind of surprised from since we haven't posted, but we are we are happy to answer questions.
- 1:05:16
- Unidentified speaker: Uh I know it's not as fun as when stocks are going to the moon every single day, but uh we will we will do our best because we we do want this to be a community where people are able to discuss what's going on across AI stocks.
- 1:05:29
- Unidentified speaker: So if you are a listener and and want to put any thoughts out there, want us to look at any additional stocks or concepts, just let us know.
- 1:05:38
- Unidentified speaker: Yes.
- 1:05:38
- Unidentified speaker: Uh Eric, thank you so much for that sentiment.
- 1:05:40
- Unidentified speaker: Um let's let let's end on that note.
- 1:05:44
- Unidentified speaker: I just wanted to add one thing because you had brought up Amazon, I thought it was a really good dovetail here.
- 1:05:48
- Unidentified speaker: Um, this is an impossibly hard uh investing is very, very challenging.
- 1:05:53
- Unidentified speaker: And most people don't appreciate.
- 1:05:54
- Unidentified speaker: I think that the majority of stocks actually underperform the index, but by getting the few major winners, you can carry the day in a portfolio and dramatically outperform.
- 1:06:03
- Unidentified speaker: So we feel the pain right there next to you, and we're not we're not here to be Pollyanish or just universally positive.
- 1:06:10
- Unidentified speaker: I'm looking at my portfolio today, and I've got Allegro Microso Microsystems, IES Holdings, Onto Innovation, Power Systems or Power Solutions, Regal Rex, Nord, Mcore, all of them are down over 30%.
- 1:06:21
- Unidentified speaker: Other ones that are not in the portfolio, Axon, um, Ad Yen, um, Netflix, these are all positions for me that are down a similar amount.
- 1:06:29
- Unidentified speaker: And yet, if you add up the losses for all of those, they are dwarfed by the gains of one single position, which is Amazon.
- 1:06:36
- Unidentified speaker: So that's not to say we should be out there, you know, we should be happy about our losses.
- 1:06:40
- Unidentified speaker: We know they don't feel good.
- 1:06:41
- Unidentified speaker: Um, but the lesson here is that you should expect in any portfolio where you're taking sufficient risk that there's going to be a lot of losers, but you don't need a lot of winners to make them completely forgotten memories and footnotes in an overall um stupendous portfolio return.
- 1:06:57
- Unidentified speaker: So not easy, dear listeners.
- 1:06:59
- Unidentified speaker: We know we, you know, we're right there with you.
- 1:07:00
- Unidentified speaker: Uh, our our money is being eviscerated, same as yours if you bought these positions, but we are still believers in this trend.
- 1:07:06
- Unidentified speaker: And uh, I think both of our portfolios long term show the value of looking through the storm clouds and you know the sort of benefits that you can achieve over longer horizons.
- 1:07:16
- Unidentified speaker: All right, Eric, let's put a pin in it at that note.
- 1:07:18
- Unidentified speaker: Uh, dear listeners, thank you so much for your time.
- 1:07:20
- Unidentified speaker: Please leave us your comments or questions, and we will see you next week.