Post-midterm market divergence highlights AI capex trajectory and Texas regulatory risks
Summary
Host Melvin analyzes the market's post-2024 midterm election reaction, noting a rally in hyperscalers like Amazon, Microsoft, and Google alongside a sell-off in semiconductor and neo-cloud stocks. He attributes this to investor fears that Democrat Gina Hinojosa's election as Texas Governor will impose stricter regulations on data center power and water usage. Melvin argues that collective AI infrastructure spending is projected to rise from $280 billion in 2024 to $1.5 trillion by 2028 due to competitive necessity. He cites strong earnings from AMD and high rental prices for H100 GPUs as evidence of sustained demand. The episode concludes by contrasting the financing risks faced by neo-cloud providers with the capital strength of hyperscalers.
Episode sections
- 0:27 – 3:16
Market Reaction to Democratic Sweep
Host introduces the episode following the 2024 midterm elections where Democrats flipped control of the House and Senate, noting initial market panic regarding AI stocks and infrastructure spending.
- 3:16 – 8:55
Texas Data Center Dominance and Political Shift
Melvin highlights Texas holding nearly a quarter of planned US data center capacity through 2030, with 63 gigawatts in projects, and discusses uncertainty introduced by the election of Democrat Gina Hinojosa as Governor.
- 8:55 – 11:40
Hyperscaler Rally vs. Semiconductor Sell-Off
Despite election fears, hyperscalers like Amazon, Microsoft, and Google rallied 5-7% while semiconductor and neo-cloud stocks faced selling pressure, as markets view hyperscalers as insulated.
- 11:40 – 22:11
Critique of Market Logic and Regulatory Reality
Melvin argues the market reaction is contradictory, noting that federal regulation is limited by veto power and Texas state changes focus on oversight rather than construction bans.
- 22:11 – 25:46
AI Capex Trajectory and Competitive Dynamics
Projected collective AI infrastructure spending is estimated to rise from $280 billion in 2024 to $1.5 trillion by 2028, driven by competitive necessity to capture market share.
- 25:46 – 30:28
Earnings Data and Hardware Demand Signals
Recent earnings from AMD, Arista Networks, and Super Micro show strong growth, with AMD reporting $13.4 billion in revenue and 45% YoY growth in data center segments.
- 30:28 – 33:24
Financing Risks and Neo-Cloud Differentiation
The AI build-out faces financing risks for neo-cloud providers relying on external debt, whereas hyperscalers raised $60 billion in bonds recently, highlighting their balance sheet strength.
- 33:24 – 36:20
Historical Post-Midterm Performance and Conclusion
Melvin cites historical data showing the S&P 500 has posted positive returns in all 19 periods following midterm elections since 1950, averaging a 14.8% gain, concluding that political uncertainty is clearing.
Assets discussed
| Asset | Kind | Stance | Context |
|---|---|---|---|
| AmazonAMZN | equity | mixed | Amazon surged 7%. |
| MicrosoftMSFT | equity | mixed | Microsoft gained six percent, Google was up like five percent, Apple all rallied at the same time. |
| AlphabetGOOGL | equity | mixed | Microsoft gained six percent, Google was up like five percent, Apple all rallied at the same time. |
| AppleAAPL | equity | neutral | Microsoft gained six percent, Google was up like five percent, Apple all rallied at the same time. |
| NVIDIANVDA | equity | mixed | Because companies like CorViev, Nebius, Iron, Nvidia, Microsoft, and several other AI companies are tied to the AI infrastructure build-out. |
| Meta PlatformsMETA | equity | mixed | Microsoft, Amazon, Google, Meta, they all have been committing extraordinary amounts of money towards AI infrastructure. |
“Stance” is how the speakers talked about the asset, not a recommendation.
Claims worth verifying
Democrats flipped control of the House and Senate in the 2024 midterm elections.
Unidentified speaker ·1:46
How to verify: Check official election results from the Federal Election Commission or major news outlets for the 2024 US Midterm Elections.
Gina Hinojosa defeated Greg Abbott to become the first Democrat Governor of Texas since Ann Richards.
Unidentified speaker ·3:16
How to verify: Verify Texas gubernatorial election results and historical records of Texas Governors.
Texas holds approximately 25% of planned US data center capacity through 2030.
Unidentified speaker ·4:29
How to verify: Review industry reports from Jeffrey or Spargo on US data center capacity forecasts.
Texas has approximately 63 gigawatts of data center capacity across existing and planned projects.
Unidentified speaker ·5:05
How to verify: Check Bloomberg charts or industry reports cited by the speaker regarding Texas data center pipeline.
Amazon surged 7%, Microsoft gained 6%, and Google was up 5% on the morning of November 5, 2024.
Unidentified speaker ·9:37
How to verify: Check historical stock price data for AMZN, MSFT, and GOOGL on November 5, 2024.
AMD reported $13.4 billion in total revenue with 45% year-over-year growth in its data center business.
Unidentified speaker ·26:09
How to verify: Review AMD's Q3 2024 earnings report released in November 2024.
Hyperscalers raised approximately $60 billion in the bond market in the week prior to the episode.
Unidentified speaker ·32:56
How to verify: Check financial news reports or bond market data for major tech issuers in early November 2024.
The S&P 500 has posted positive returns in all 19 periods following midterm elections since 1950, averaging a 14.8% gain.
Unidentified speaker ·33:44
How to verify: Analyze historical S&P 500 performance data relative to midterm election dates from 1950 to present.
Projected collective AI infrastructure spending by major players will reach $1.5 trillion by 2028.
Unidentified speaker ·23:02
How to verify: Review analyst estimates or company guidance regarding AI capex projections for 2024-2028.
Meta raised $58 billion in investor orders for $15 billion in dollar-denominated bonds.
Unidentified speaker ·32:12
How to verify: Check Meta's financial filings or news reports regarding their bond issuance in 2024.
Do your own research
- Verify Texas data center capacity projections and Gina Hinojosa's specific policy proposals via Texas state government filings and news archives.
- Check AMD's latest 10-Q or 10-K filing for detailed data center revenue breakdowns and growth rates.
- Review hyperscaler bond issuance records to confirm the $60 billion fundraising figure and compare it to neo-cloud provider debt structures.
- Analyze GPU rental market rates for H100 and other models using cloud marketplace data or industry reports to assess supply/demand dynamics.
- Review historical S&P 500 performance data following midterm elections from 1950 to present to verify the 14.8% average gain claim.
- Assess the regulatory landscape for data centers in Texas, including water usage and power consumption rules, through public utility commission documents.
Full transcriptexpand
- 0:00
- Unidentified speaker: This is essentially a buy the dip moment for me, and I will be buying some of my favorite companies in memory, optical networking, neo clouds, etc.
- 0:08
- Unidentified speaker: And if you're an investor, volatility, as I always say, brings opportunity.
- 0:12
- Unidentified speaker: Hello everyone.
- 0:13
- Unidentified speaker: Welcome to Milk Road Stocks, the daily market show that is here to remind you not to refresh your portfolio this morning.
- 0:18
- Unidentified speaker: Uh look, it's ugly out there.
- 0:20
- Unidentified speaker: My name is John.
- 0:20
- Unidentified speaker: This is Milk Road Stocks.
- 0:22
- Unidentified speaker: Today is Thursday, November 5th, and we are recording this on Wednesday, November 4th.
- 0:27
- Unidentified speaker: Just a few days ago, Wall Street was selling the hyperscalers for spending too much on AI.
- 0:30
- Unidentified speaker: Now the Democrats have flipped the House and the Senate and taking control of Congress, and everyone is panicking about what this means for AI stocks, for the AI trade and for every level of that trade.
- 0:40
- Unidentified speaker: Are these companies spending too much?
- 0:41
- Unidentified speaker: Are they about to stop spending?
- 0:43
- Unidentified speaker: Are they going to be forced to stop spending?
- 0:45
- Unidentified speaker: Today, Melvin, our leading portfolio manager, is on the show to explain why he thinks the market is overreacting here.
- 0:50
- Unidentified speaker: And I hope that he is right.
- 0:51
- Unidentified speaker: Uh, he's saying that Texas holds nearly a quarter of the planned US data center capacity, just by one estimate.
- 0:56
- Unidentified speaker: Ohio, Georgia also have a lot as well.
- 0:58
- Unidentified speaker: But he's gonna go into what this means, why he thinks that you should be paying attention here, and he also tells us whether he is actually buying this dip, which is what everybody wants to know right now on a day like today.
- 1:06
- Unidentified speaker: So he's gonna tell all of us what he's thinking here.
- 1:09
- Unidentified speaker: If you want to see his portfolio, his actual trades that he's making, you can join Milk Road Pro.
- 1:13
- Unidentified speaker: The link is in the description, it's just a dollar sign up there.
- 1:15
- Unidentified speaker: Uh, and we'll see you there.
- 1:16
- Unidentified speaker: So without further ado, enjoy the show.
- 1:17
- Unidentified speaker: This one's gonna be a really good episode.
- 1:19
- Unidentified speaker: I think it's really timely to be talking about this today.
- 1:20
- Unidentified speaker: Uh, so I hope we enjoy this, and thanks for being here.
- 1:23
- Unidentified speaker: Wow, what a brutal market this has been.
- 1:26
- Unidentified speaker: Um, last night, Americans went to the poll for what turned out to be the most one of the most consequential midterms in the recent history.
- 1:35
- Unidentified speaker: And uh looking at the results, I think it's fair to say things went significantly better for Democrats than many investors, including myself, were anticipating just a few months ago.
- 1:46
- Unidentified speaker: Democrats managed to flip both control of the House and the Senate, taking away uh Republicans' control of Congress and setting up what could be two years of political gridlock in Washington.
- 2:00
- Unidentified speaker: They also secured several major victories at the state level, including races in places that historically have been extremely difficult for Democrats to win.
- 2:11
- Unidentified speaker: Um, in the Senate, Democrats roadly flipped seven Republican held seats, including Texas, Ohio, Kansas, Iowa, uh, Maine, North Carolina, and South Carolina.
- 2:24
- Unidentified speaker: And according to the figures, um, Democrats are are still sitting at around 54 right now, um, compared to 44.
- 2:33
- Unidentified speaker: And I believe Mississippi and Alaska are still being counted.
- 2:37
- Unidentified speaker: Um, meanwhile, Democrats also crossed the threshold for a 229 seat to take control of the House of Representatives, um, although the final you know makeup of the Congress is still being determined, but it's safe to say they they came and they conquered, conquer really.
- 2:55
- Unidentified speaker: It's been a blue tsunami um, you know, all of yesterday and today.
- 2:59
- Unidentified speaker: So why is that important for us?
- 3:01
- Unidentified speaker: Because for investors like us, especially invested in AI and semiconductor data centers, energy infrastructure, I think one of the most sequential and consequential results of the entire night came out of Texas, right?
- 3:16
- Unidentified speaker: Gina defeated um Governor Greg Abbott, becoming the first Democrat governor of Texas since Ann Richards.
- 3:26
- Unidentified speaker: Um, James also defeated Ken Paxton in the Senate race, marking another major victory and a major shift in the Texas state level as well.
- 3:37
- Unidentified speaker: Now, why does Texas matter so much for AI?
- 3:40
- Unidentified speaker: Because Texas has become one of the most important locations for data center development in the United States.
- 3:46
- Unidentified speaker: The state has relatively abundant land, enormous energy um industry, and a business environment that historically attracted some of the largest and infrastructure companies in the world.
- 3:59
- Unidentified speaker: So, and we saw this massive shift, you know, back in like the great migration happened back in like 2020-ish, um, I would say, and uh a lot of companies just have been just moving down there um because of how far friendly they are.
- 4:14
- Unidentified speaker: And um companies have announced, like these data center companies have announced like massive projects across Texas, and uh a substantial portion of that um country's planned data center expansion is obviously in Texas.
- 4:29
- Unidentified speaker: So, according to Jeffrey estimates, um it roughly a quarter of planned US center data center capacity coming online through 2030 is located in Texas.
- 4:40
- Unidentified speaker: While Spargo, on the other hand, also estimated that state accounts for roughly 27% of existing US data center capacity.
- 4:48
- Unidentified speaker: So we're talking about a state that could play an enormous role in determining how quickly the next generation of AI infrastructure gets developed.
- 4:57
- Unidentified speaker: And when you look at this Bloomberg chart that I just put up, you can just see how enormous Texas pipeline has become.
- 5:05
- Unidentified speaker: Texas has approximately 63 gigawatts of data center capacity across existing and planned projects, especially significantly more than Georgia and Virginia, which are both around like 40 gigawatts.
- 5:19
- Unidentified speaker: But here's what I find interesting a substantial portion of Texas pipeline is still in the early stages of development, and much of that capacity is being pursued by first-time developers rather than established operators.
- 5:33
- Unidentified speaker: Because that matters because these projects haven't necessarily secured everything, they need to move forward.
- 5:41
- Unidentified speaker: Some still require additional financing, electricity agreements, permitting approvals, yada yada yada.
- 5:49
- Unidentified speaker: And if Texas starts to begin implementing stricter standards around power consumption and water usage and all the other community guidelines and protections, these those early stages that I'm talking about could face the greatest uncertainty.
- 6:05
- Unidentified speaker: And uh that's essentially what the what the issue is.
- 6:08
- Unidentified speaker: And this chart also reinforces my broader argument that Texas already has an enormous amount of AI plant infrastructure investment, and the demand driving those projects isn't going to suddenly disappear because of an election.
- 6:23
- Unidentified speaker: Some developments could face delays, others might become more expensive, and certain projects may never get built.
- 6:30
- Unidentified speaker: I think it's far more likely that we see companies adopt to a change of regulatory environment rather than abandon their expansion plans altogether.
- 6:41
- Unidentified speaker: And that brings me to Gina Hinosa's proposed policy because I think investors are overlooking what she's actually asking developers to do.
- 6:53
- Unidentified speaker: So Hinosa has been particularly outspoken about the need for a stricter oversight of new data center construction, including calling temporary ban on data centers while lawmakers establish stronger protection for local communities.
- 7:09
- Unidentified speaker: Across the countries, communities have become increasingly concerned about the amount of electricity and water these facilities consume, the pressure they place on local infrastructure, and whether you know ordinary households will ultimately be responsible for paying some of the costs associated with construction.
- 7:30
- Unidentified speaker: And we're starting to see much broader political conversation about the AI infrastructure, not necessarily because people want to oppose AI, but because they want to understand what these enormous facilities will mean for their communities.
- 7:46
- Unidentified speaker: Are electricity bills going to increase?
- 7:48
- Unidentified speaker: Is there enough water?
- 7:50
- Unidentified speaker: Who pays for new transmission lines?
- 7:52
- Unidentified speaker: Are these projects creating permanent jobs, or are most of these jobs gonna get replaced once the construction finishes?
- 8:01
- Unidentified speaker: You know, these are legitimate questions and concerns.
- 8:05
- Unidentified speaker: But I think the market may be mistaking, mistaking this by interpreting growing political opposition to certain data center projects as evidence that the entire AI infrastructure cycle is about to slow down.
- 8:19
- Unidentified speaker: And and simply that's just isn't true, in my opinion.
- 8:23
- Unidentified speaker: And that brings me to what happened in the stock market uh this morning or yesterday when you when you see this video.
- 8:31
- Unidentified speaker: We're gonna pause for a second.
- 8:32
- Unidentified speaker: The AI trade is changing.
- 8:34
- Unidentified speaker: The bottleneck stocks are either running out of steam or getting ready for another leg higher, while the application layer is finally starting to show some real profits.
- 8:42
- Unidentified speaker: Our analysts were early to the last wave, and now they are buying for the next one.
- 8:46
- Unidentified speaker: Catch your next call inside Milk Road Pro for just a dollar at the link below.
- 8:50
- Unidentified speaker: My question is Is Wall Street abandoning semiconductors for stocks for hyperscalers?
- 8:55
- Unidentified speaker: Because look at this more yesterday's uh reaction, like morning reaction.
- 9:02
- Unidentified speaker: So, to give you some context, following last night's or following earnings election night results, NASDAQ was down approximately 2.4 percent.
- 9:12
- Unidentified speaker: And it's investors try to figure out what a democratic sweep could mean for the technology sector.
- 9:18
- Unidentified speaker: But when the market opened this morning, something very interesting happened, and that is it opened green, super green.
- 9:28
- Unidentified speaker: Um, instead of seeing a broad sell-off across technology, we saw how actually an aggressive rally in some of the largest hyperscalers.
- 9:37
- Unidentified speaker: Amazon surged 7%.
- 9:39
- Unidentified speaker: I think that's their second biggest, largest run-up in the last two years.
- 9:44
- Unidentified speaker: Microsoft gained six percent, Google was up like five percent, Apple all rallied at the same time.
- 9:52
- Unidentified speaker: Several semiconductor or all of the semiconductor trade, I would say, like memory, neo clouds, opticals, they're under severe selling pressure.
- 10:04
- Unidentified speaker: Um, why?
- 10:06
- Unidentified speaker: Because companies like CorViev, Nebius, Iron, Nvidia, Microsoft, and several other AI companies are tied to the AI infrastructure build-out.
- 10:16
- Unidentified speaker: We're we're trading considerably weaker in the hyper and then the hyperscaler.
- 10:21
- Unidentified speaker: And I think this kind of tells us about how the investors are kind of interpreting this election.
- 10:28
- Unidentified speaker: The market seems to believe that companies like Google, Amazon, and Microsoft are relatively insulated from potential regulatory changes because they already operate enormous cloud business and they operate and they generate substantial like cash flow from their other lines of business.
- 10:49
- Unidentified speaker: Meanwhile, companies whose growth heavily depends on building new facilities or securing additional secure um electricity and raising money to fund that uh finance or fund that build out could face more uncertainty.
- 11:05
- Unidentified speaker: And that's what the market is saying right now.
- 11:08
- Unidentified speaker: And to be fair, I think there is some logic behind that, because if new data center construction becomes more difficult, existing computing capacity becomes more valuable.
- 11:18
- Unidentified speaker: Companies that have large access to electricity, buildings, networking, large GPU clusters could benefit from that scarcity.
- 11:28
- Unidentified speaker: You know, that's the flip side of this.
- 11:30
- Unidentified speaker: But I think investors are getting way, way too ahead of themselves with this semiconductor sell-off because there is a pretty big contradiction in how the market is reacting.
- 11:40
- Unidentified speaker: Because think about it, just days ago during the earnings, right, investors were selling these same hyperscalers because they were concerned about their enormous capex.
- 11:51
- Unidentified speaker: Microsoft, Amazon, Google, Meta, they all have been committing extraordinary amounts of money towards AI infrastructure.
- 11:59
- Unidentified speaker: We all know that.
- 12:00
- Unidentified speaker: And investors for the longest were bringing to question whether those investments could generate sufficient returns.
- 12:09
- Unidentified speaker: And we were just hearing concerns about negative cash flow, rising depreciation expenses, additional debt issuance, and whether companies are spending too aggressively on AI.
- 12:22
- Unidentified speaker: Now, following the election, many of those stocks are rallying because investors believe the new regulations could potentially slow down the infrastructure companies supplying them.
- 12:33
- Unidentified speaker: So which is it?
- 12:34
- Unidentified speaker: Are investors worried that hyperscalers are spending too much money, or are they worry that the same companies are suddenly going to stop spending?
- 12:43
- Unidentified speaker: Because if the AI build out genuinely slows down to materially hurt the semiconductor demand, I don't see that's I don't see how that would be particularly bullish for cloud businesses that depends on these on the broader infrastructure.
- 12:59
- Unidentified speaker: Because look at these revenue from cloud business over the last couple of years from AWS, Azure, Google Cloud, they have been growing tremendously.
- 13:09
- Unidentified speaker: And what is the driver of all this?
- 13:11
- Unidentified speaker: It's because of AI.
- 13:13
- Unidentified speaker: So what I don't understand about this market reaction is that if market believes Amazon and all these cloud providers are going to continue to grow their business, why is it that they're pricing in a major slowdown for companies supplying the compute infrastructures those companies depend on?
- 13:31
- Unidentified speaker: So that makes no sense.
- 13:33
- Unidentified speaker: Now I understand why investors might prefer companies with a you know stronger balance sheet during this period of uncertainty.
- 13:40
- Unidentified speaker: That's completely reasonable.
- 13:42
- Unidentified speaker: But there's a difference between temporarily favoring one part of the AI ecosystem over another and concluding that the long-term growth prospects of semiconductor companies have you know suddenly disappeared.
- 13:56
- Unidentified speaker: I think much of what we're seeing today is a temporary risk-off rotation following the election because nobody knows where we go from here rather than like some fundamental changes in the AI investment cycle.
- 14:11
- Unidentified speaker: And it's also worth remembering that semiconductor stocks have already been under heavy pressure even before yesterday's election.
- 14:21
- Unidentified speaker: The Philadelphia Semiconductor Index have experienced a massive, massive correction over the last I would say month or so.
- 14:30
- Unidentified speaker: And uh so some some of the selling may be continuation of those of an existing correction, with the election simply providing another reason for investors to reduce exposure.
- 14:44
- Unidentified speaker: One morning of price action doesn't establish, in my opinion, a permanent change in the market leadership.
- 14:52
- Unidentified speaker: I wouldn't be surprised if some of these semiconductor names, companies getting sold today, recover much, much sooner than investors actually expect.
- 15:01
- Unidentified speaker: But to understand why I feel that way, I think we need to look more closely at what the election actually changes, right?
- 15:09
- Unidentified speaker: What does that actually mean?
- 15:10
- Unidentified speaker: Let's start with Washington, because I think there's a reason why a democratic democratic sweep doesn't necessarily necessarily translate into sweeping new restrictions for AI.
- 15:23
- Unidentified speaker: Even though Democrats now control Congress, Donald Trump is still the president.
- 15:28
- Unidentified speaker: That means any major federal regulation or legislation would gener would generally need either his signature or enough congressional support to override a presidential veto.
- 15:42
- Unidentified speaker: And democrats don't have two-thirds majority to override a veto veto on their own.
- 15:49
- Unidentified speaker: So while Congress can introduce new legislations or hold hearings and investigate technology technology companies, it doesn't mean it can simply pass a nationwide ban on data center construction.
- 16:01
- Unidentified speaker: That is not going to happen at all.
- 16:02
- Unidentified speaker: And I think that is somewhat overlaw overlooked in the initial reaction because JP Morgan and Goldman Sachs have both discussed the possibility that a Democratic Congress would actually increase regulatory scrutiny without necessarily changing the near-term trajectory of AI infrastructure investment.
- 16:22
- Unidentified speaker: In other words, the political headlines might change significantly, but the actual legislative side of things could remain pretty much the same.
- 16:33
- Unidentified speaker: Now, I'm not saying that like election is irrelevant.
- 16:36
- Unidentified speaker: Democrats can absolutely use their control of like congressional committees to investigate major technology companies and question executives about their spending, scrutinize everything from electricity to government subsidies, but they could also put pressure on federal agencies over national, you know, environmental standards, labor regulations.
- 16:59
- Unidentified speaker: But I think the more immediate risk is actually happening at the late, late state and local levels.
- 17:04
- Unidentified speaker: And that's where we need to talk about Texas.
- 17:07
- Unidentified speaker: Let's you know double-click on Texas a little bit.
- 17:10
- Unidentified speaker: Because we know Gina's victory has obviously created uncertainty for companies developing AI infrastructure in Texas.
- 17:18
- Unidentified speaker: And when you look at the planned investment in the state, it's you know, as I mentioned earlier, it's it's understandable why.
- 17:24
- Unidentified speaker: According to Jeffreys, he believes that 13.3 gigawatts of um announced Texas Center project could face uncertainty.
- 17:34
- Unidentified speaker: And uh Bank of America is kind of essentially saying the same thing.
- 17:38
- Unidentified speaker: We're talking about hundreds of you know, gigawatts of proposed electricity demand.
- 17:43
- Unidentified speaker: But although it's important to remember that not every proposed project will actually get built, we all know that.
- 17:49
- Unidentified speaker: Still, the scale of these plans over in Texas is extraordinary, and the companies with the most concentrated exposure to Texas could face greater regulatory uncertainty than companies with like very geographically diverse, diversified operations, Wink Wink and Nebbias.
- 18:07
- Unidentified speaker: Uh, but the market doesn't care, unfortunately, right now.
- 18:09
- Unidentified speaker: It's selling off everything because it does not matter.
- 18:13
- Unidentified speaker: Like if you look at names like Iron, Galaxy, Cypher, Riot, Core Weave, they're massively connected to the Texas infrastructure development.
- 18:24
- Unidentified speaker: Meanwhile, companies like Nebias and Applied Digital and TerraWolf, you know, have different geographical footprints, which can make the implication of tighter regulations, Texas regulation, very different for each business.
- 18:37
- Unidentified speaker: But if you look at actually look at more closely at Hinoha's like actual policy proposal, I think the market is like severely overreacting to this because there's a difference between saying that the new data center construction needs better oversight and saying that companies can no longer build data center um data centers.
- 19:00
- Unidentified speaker: And those are not the same thing because I was looking at Hinoha's own website, and the priorities she outlined are quite specific.
- 19:09
- Unidentified speaker: She called for stronger labor protection, requiring developers to employ skilled American union workers and giving local communities a greater voice in whether these facilities should be built near them.
- 19:24
- Unidentified speaker: She also proposed greater transparency, including restrictions on non-disclosure agreements that can prevent residents from understanding the details of what's actually being developed in the communities.
- 19:36
- Unidentified speaker: And two of her biggest priorities relate to electricity and water.
- 19:41
- Unidentified speaker: She wants data center developers to demonstrate that surrounding households won't face higher electricity or water costs because of new projects, you know.
- 19:53
- Unidentified speaker: And she wants developers to essentially take more responsibility to financing their own infrastructure rather than depending on tax paying dollars or subsidies.
- 20:04
- Unidentified speaker: And finally, she proposed stronger protections for you know water, local water supplies, and particularly in situations where large industrial water contracts would threaten the resources available to residents.
- 20:19
- Unidentified speaker: Is any of this crazy?
- 20:21
- Unidentified speaker: Like, I don't think what she's asking is crazy at all.
- 20:25
- Unidentified speaker: I think these are pretty reasonable asks.
- 20:28
- Unidentified speaker: Like, if you want to build data centers in our community, you have to work with the community a little bit more better.
- 20:36
- Unidentified speaker: You have to demonstrate that your projects won't fit place an unfair financial burden on the people living there.
- 20:44
- Unidentified speaker: I mean, that is what everyone should realistically do.
- 20:49
- Unidentified speaker: But some of these media outlets like to spin this, you know, spin these um takes as oh, they're shutting data centers down.
- 20:57
- Unidentified speaker: No, I think data centers provide an enormous amount of jobs in the community.
- 21:03
- Unidentified speaker: It's it's growing, like it's the largest part part of our GDP like growth as of now.
- 21:11
- Unidentified speaker: If the data center build out gets slower, I think we're gonna severely because severely be um potentially going towards a negative GDP growth.
- 21:23
- Unidentified speaker: And if you have two two consecutive months of negative GDP growth, congratulations, you're on you're on recession right now.
- 21:31
- Unidentified speaker: You have a recession in your hands.
- 21:33
- Unidentified speaker: So nobody wants that.
- 21:34
- Unidentified speaker: I I don't think Democrats or Republicans want a recession in their hands because now Democrats and Republicans will point fingers at each other and say, oh, you cost it, you cost it, but in reality, both of them will cost this if if if the build out don't continue.
- 21:54
- Unidentified speaker: So it is essential for this build out um to take place, and I think companies need to work with these communities to build these data centers out.
- 22:04
- Unidentified speaker: Now, here's why I don't think the AI build out is gonna slow down anytime soon.
- 22:11
- Unidentified speaker: Because AI build out is already well underway.
- 22:14
- Unidentified speaker: We all know that.
- 22:16
- Unidentified speaker: Amazon, Google, Microsoft, Meta collectively are spending billions, okay, billions of dollars.
- 22:23
- Unidentified speaker: Look at these, look at these three capex estimates.
- 22:26
- Unidentified speaker: You know, I think this just got revised just this week alone or last week.
- 22:32
- Unidentified speaker: Meta, Google, Amazon, Oracle, Core Weaver Nebia, SpaceX could increase from approximately two or did increase from approximately 280 billion in 2024 to 951 billion in 2026.
- 22:47
- Unidentified speaker: By 2027, that projected figure is gonna reach 1.37 trillion.
- 22:55
- Unidentified speaker: Okay, and by 2028, we're talking about 1.5 trillion going into these data centers.
- 23:02
- Unidentified speaker: Now, I think there is a reasonable possibility that spending projects could continue moving higher.
- 23:08
- Unidentified speaker: Why?
- 23:09
- Unidentified speaker: Because we're still relatively early in the development of AI.
- 23:14
- Unidentified speaker: Think about how much AI has changed over the past few years.
- 23:18
- Unidentified speaker: We have essentially gone from large language models that could answer questions and generate text to increasingly sophisticated AI agents that runs for days.
- 23:30
- Unidentified speaker: And every new generation of the system requires enormous computing resources.
- 23:36
- Unidentified speaker: And as more businesses integrate AI into their everyday operations, the computing power required to serve those applications could increase trim significantly.
- 23:47
- Unidentified speaker: And we're still so early, that's what I'm saying.
- 23:50
- Unidentified speaker: We're probably in the second or the third inning of this AI infrastructure build out.
- 23:55
- Unidentified speaker: And uh as investors, the profits may not come till the sixth, seventh, or the eighth, or even the ninth inning, the majority of the profits.
- 24:03
- Unidentified speaker: We're starting to see the profits come in, but the infrastructure has to get built up before before the profits come in.
- 24:10
- Unidentified speaker: So another reason I think that we're not these this CapEx number is gonna go higher is because competition.
- 24:17
- Unidentified speaker: We all know hyperscalers are competing against each other.
- 24:21
- Unidentified speaker: If Microsoft decides to significantly reduce his AI infrastructure while Amazon continues to expand AWS, Microsoft essentially risks losing customers that need access to you know additional computing capacity to Amazon.
- 24:36
- Unidentified speaker: If Google slows down its investment in AI infrastructure while Microsoft and OpenAI keep pushing forward, Google risks losing ground in enterprise AI and cloud computing.
- 24:48
- Unidentified speaker: And if Meta decides to substitute you know reduce spending on AI models, somebody else is gonna take up take their position.
- 24:56
- Unidentified speaker: These companies understand that.
- 25:10
- Unidentified speaker: So the TAM is a lot here.
- 25:14
- Unidentified speaker: So I don't think this CapEx is gonna slow down.
- 25:18
- Unidentified speaker: Yes, it might slow down like in terms of like the dollar amount, because it it's just a large, it's just a numbers game.
- 25:26
- Unidentified speaker: Uh it's gonna get smaller, right?
- 25:28
- Unidentified speaker: It can't keep going higher every year.
- 25:31
- Unidentified speaker: But what you have to understand is like companies are still spending billions and trillions of dollars to build out this AI infrastructure, and as I said, revenue is gonna be a big deciding factor on all this.
- 25:46
- Unidentified speaker: Revenue hasn't really come in yet.
- 25:49
- Unidentified speaker: Okay, we're still early, so it will eventually start to come in because yesterday alone we had a couple great earnings from companies like AMD.
- 26:00
- Unidentified speaker: AMD reported significant growth in their data center business.
- 26:05
- Unidentified speaker: I think they were at one point they were up like six or seven percent after hours.
- 26:09
- Unidentified speaker: AMD generated $13.4 billion in total revenue, representing a 45% year-over-year growth in this data center business, generated approximately $8.5 billion, up 97% from previous last year.
- 26:25
- Unidentified speaker: It doubled in 12 months, and AMD isn't the only one benefiting from this expansion.
- 26:31
- Unidentified speaker: Arista Networks has been report reported strong demand for high performance networking equipment.
- 26:37
- Unidentified speaker: They also had their earnings yesterday.
- 26:39
- Unidentified speaker: Massive beat.
- 26:40
- Unidentified speaker: Supermicro, another name, experienced significant growth tied to AI server um developments.
- 26:47
- Unidentified speaker: There's so many earnings that came out yesterday, all pointed that AI demand is stronger than ever before.
- 26:55
- Unidentified speaker: And of course, the stocks are down or a little bit flat because of weakness in semis, and that's that's fine.
- 27:02
- Unidentified speaker: And even look at GPU rentals, okay.
- 27:05
- Unidentified speaker: Nvidia B200 and H100 GPUs have had significantly like went up during the year and it still continue to go up.
- 27:14
- Unidentified speaker: And H100 is isn't even the NVIDIA's newest generation.
- 27:18
- Unidentified speaker: It's it's so old.
- 27:20
- Unidentified speaker: If like these older computing hardware is still commanding rental high prices, suggests that there may be more underlying demand for capacity than compute is available.
- 27:31
- Unidentified speaker: Okay, I I think you guys need to understand it.
- 27:34
- Unidentified speaker: Understand this.
- 27:35
- Unidentified speaker: This is compute need for compute is going up, it's not slowing down.
- 27:40
- Unidentified speaker: Now, let's let me bring it back to the semiconductor sell-off because I think investors, this is where investors could be missing an opportunity.
- 27:48
- Unidentified speaker: I think the market seems to be assuming that tighter restrictions on data center will translate into lower demand for semiconductors.
- 27:56
- Unidentified speaker: But I don't think it's true.
- 27:57
- Unidentified speaker: Even if data center approvals become difficult, that doesn't mean semiconductor demand disappears.
- 28:04
- Unidentified speaker: Existing this facilities still need more powerful GPUs, bandwidth memory, network equipment, yada yada yada.
- 28:12
- Unidentified speaker: And if developers can build any new facilities as they were originally planned, they have a greater incentive to maximize the computing capacity of the infrastructure they already have.
- 28:23
- Unidentified speaker: Think about it.
- 28:24
- Unidentified speaker: If Amazon or Microsoft already has an operational data center with access to electricity, cooling systems, yada, yada, yada, that facility becomes increasingly valuable if new construction becomes more difficult.
- 28:37
- Unidentified speaker: So if you're a neo cloud provider that has these available capacity, your your spot price just went out because if there is less compute to go around, who benefits?
- 28:51
- Unidentified speaker: People do that have computing capacity.
- 28:54
- Unidentified speaker: And now look at this chart.
- 28:56
- Unidentified speaker: This is the AI stack right here.
- 28:58
- Unidentified speaker: Broadcom companies like Nvidia, AMD, Broadcom, Credo, Lite, AAOI could continue benefiting.
- 29:07
- Unidentified speaker: Let's start with NVIDIA.
- 29:08
- Unidentified speaker: We all know NVIDIA remains one of the most suppliers in the entire ecosystem, regardless of whether Microsoft or Google or another company starts to slow down the AI race.
- 29:26
- Unidentified speaker: Brockcom is another company, I think, Mar uh particularly interesting because it's exposure to custom silicon and networking.
- 29:33
- Unidentified speaker: Marvell, Marvell is thinking about is gonna acquire light, and they're gonna that's gonna create a massive moat in their business as well.
- 29:41
- Unidentified speaker: So these are all some of the companies that are benefiting from the AI infrastructure demand.
- 29:47
- Unidentified speaker: And then as I mentioned, like Credo, Lamentum, Apply Electronics are like helping remove one of the bottlenecks in the industry, which is like networking.
- 29:57
- Unidentified speaker: And then we have stuff like memory, companies like Micron, SK Heinick, Samsung has you know benefited from growing demand for high bandwidth memory and other memory products.
- 30:06
- Unidentified speaker: So you get it.
- 30:08
- Unidentified speaker: You have to understand there's so many bottlenecks right now, like power, memory, substrats, um, CPU, GPU, like there's the entire supply chain is constrained right now.
- 30:20
- Unidentified speaker: And you have to understand demand for all this is only going higher.
- 30:26
- Unidentified speaker: And that brings me to the broader opportunity.
- 30:28
- Unidentified speaker: I think the investors are still treating the AI infrastructure as though every company faces the same regulatory and financial financing risks, but the reality is much more complicated.
- 30:40
- Unidentified speaker: A semiconductor selling into multiple hyperscalers is very different from a neo cloud, a or a lower tier, I should say lower tier neo cloud that borrows billions of dollars to fund build a new facility in Texas.
- 30:54
- Unidentified speaker: And even within the neo cloud industry, I've told you guys this core weeb and um Nebius still remains the strongest.
- 31:02
- Unidentified speaker: They can actually get the funding.
- 31:04
- Unidentified speaker: And so I think this sell-off could actually create some very attractive points.
- 31:09
- Unidentified speaker: And um, but I would definitely focus on companies with strong demand, real customers, reasonable financing needs, and a credible path flow to generating cash flow.
- 31:21
- Unidentified speaker: I think there's a lot of opportunities that you can buy here.
- 31:24
- Unidentified speaker: This is a buy the dip moment.
- 31:26
- Unidentified speaker: I'm gonna buy some companies here.
- 31:29
- Unidentified speaker: And if you want to know what companies I'm buying, you know, you can just subscribe to Milk Road for just it's just a dollar.
- 31:34
- Unidentified speaker: You can come check it out.
- 31:35
- Unidentified speaker: Now, I don't want to sound like there's no risk to the AI build out.
- 31:38
- Unidentified speaker: There definitely is.
- 31:39
- Unidentified speaker: I think one of the biggest reasons risk is financing because for the last several years, um, much of the AI infrastructure build out um has been funded by some of the most profitable companies in the world.
- 31:53
- Unidentified speaker: Companies like Microsoft, Google, Meta now sing now give them that gives them a significant advantage over smaller infrastructure companies that heavily depend on external financing, right?
- 32:07
- Unidentified speaker: Now, but investors are now starting to pay close attention because take Meta, for example.
- 32:12
- Unidentified speaker: Meta received as much as $58 billion in investor investor order for approximately $15 billion of uh denom dollar denominated bonds.
- 32:22
- Unidentified speaker: That's an enormous amount of demand, and companies are still willing to fund these hyperscalers, and that's that's the reality.
- 32:29
- Unidentified speaker: It's just harder for these other businesses to get funding if you don't have a strong cash flow or a real good deal, like these NeoClouds, like a hyperscaler deal that they can go to the bank and get a loan for.
- 32:44
- Unidentified speaker: Now look at this, look at this chart I just put up this week alone, okay, or in a in uh about a week, hyperscalers raise as much as $60 billion bond market.
- 32:56
- Unidentified speaker: That's enormous.
- 33:08
- Unidentified speaker: So, where does this all leave us?
- 33:11
- Unidentified speaker: Right?
- 33:11
- Unidentified speaker: I came to the election believing that the AI infrastructure was build out was bigger than politics, and looking at everything over the last 24 hours, and I still believe that.
- 33:22
- Unidentified speaker: Just look at this chart.
- 33:24
- Unidentified speaker: The midterm streak in the past 19 periods has been extremely remarkable.
- 33:30
- Unidentified speaker: Since 1950, SP 500 has developed possible returns, positive returns in all 19 number to April periods following a midterm election with an average gain of 14.8 percent.
- 33:44
- Unidentified speaker: That's a perfect 19 for 19 record for over more than 70 years.
- 33:48
- Unidentified speaker: And what's interesting is that even during periods of economic uncertainty and political gridlock and market volatility, as you're seeing, stocks have consistently moved higher during this six-window, six-month window.
- 34:01
- Unidentified speaker: Now, I'm not saying history guarantees another rally, especially how much the market has already run, and we're in new, like uncharted territory with the AI build out, but I think it's worth keeping in mind as investors react to these invest election results.
- 34:16
- Unidentified speaker: And the political uncertainty is starting to clear right now.
- 34:20
- Unidentified speaker: And I think the huge reason that Democrats or Republicans lost is because of the war, and JD ran so they actually came out and admitted that.
- 34:29
- Unidentified speaker: So it is in their best interest for them to end the war in Iran as as soon as possible because if it if they don't, if this war still drags on, Republicans have no shot at winning in 2028.
- 34:43
- Unidentified speaker: So, and they don't obviously no one wants that um from their party.
- 34:47
- Unidentified speaker: So I think it's in their best interest for us to have really good market gains and really good economy.
- 34:54
- Unidentified speaker: So they're gonna do everything in their power over the next course of a year and a half to really pump up this economy, and that's why I keep saying this is like a buy the dip opportunity here, because there's so many good companies that are on sale right now on the semiconductor side, and I have a strong, very, very strong feeling that we're we're gonna continue to go higher, and this is essentially a buy the dip moment for me.
- 35:20
- Unidentified speaker: And I will be buying some of my favorite companies in memory, optical networking, um, neo clouds, etc.
- 35:27
- Unidentified speaker: And if you're an investor, you know, volatility, as I always say, brings opportunity.
- 35:32
- Unidentified speaker: You have to take advantage of these moments.
- 35:36
- Unidentified speaker: And uh, if you do, the market will reward you.
- 35:40
- Unidentified speaker: Just zoom out, think about the bigger picture, think about where AI is going, and that's essentially all you have to do here.
- 35:48
- Unidentified speaker: Thank you.
- 35:48
- Unidentified speaker: Thanks for listening to Milk Road.
- 35:50
- Unidentified speaker: If you enjoyed the show, make sure you like and subscribe.
- 35:52
- Unidentified speaker: And if you're struggling to find winners in the market, that's exactly what Milk Road Pro is built for.
- 35:57
- Unidentified speaker: Our analysts have called some of the biggest winners early, and Pro lets you see what they're buying next, every trade they make, and the research behind every position.
- 36:05
- Unidentified speaker: Check out Milk Road Pro at the link below.
- 36:07
- Unidentified speaker: Everything you hear on Milk Road is for informational purposes only.
- 36:10
- Unidentified speaker: These are our personal opinions, not financial advice.
- 36:13
- Unidentified speaker: And we may own some of the investments we talk about.
- 36:15
- Unidentified speaker: Always do your own research and make the decisions that are right for you.
- 36:18
- Unidentified speaker: See you next time.
More from Milk Road Stocks
Nvidia Is Down 30%. How Much Was Jensen Huang Worth?
Episode discussion: Nvidia Stock Drops 30% After Jensen Huang's Departure Amid Political Pressure
Speakers’ claims: John Gillen and Vincent discuss Nvidia's 30% stock decline and $1 trillion market cap loss following Jensen Huang's departure as CEO after 33 years. Vincent attributes the exit to political pressure from both Republican and Democratic officials regarding Huang's ties to Alice Huang, framing the leadership change as collateral damage in pre-midterm politics. The speakers analyze the operational risks of losing a founder-led CEO, comparing Nvidia's structure to Tesla's dependency on Elon Musk, particularly regarding innovation speed and pricing power. While acknowledging the leadership vacuum, Vincent cites analyst Tom Lee's argument that Nvidia's deep engineering talent and asset-light model support a 'buy the dip' thesis. The episode also covers geopolitical risks, including the RACE Act's impact on Chinese chip innovation and the potential for a rival AI stack.
One Earnings Report Just Flipped the Entire AI Stock Market
Episode discussion: CoreWeave earnings flip AI market sentiment from survival to growth
Speakers’ claims: Speaker 0 reports that CoreWeave's Q3 results, featuring a 3% revenue beat and significant margin improvements, shifted market narrative from cash burn concerns to growth potential. The speaker highlights CoreWeave's $132.5 billion backlog and doubled contract pricing, noting that switched-on power, not demand, is the primary bottleneck. Bearish concerns regarding $46.9 billion in debt and $7.5 billion quarterly cash burn are acknowledged but offset by management's 2028 profitability target and customer prepayments. The earnings triggered rallies in AI infrastructure stocks like Nebius, IREN, and Oracle, with the speaker projecting strong results for Nebius. The episode concludes with anticipation for Anthropic's IPO and a disclaimer that the content is not financial advice.