The Midterms Hit AI: Data Center Pushback, a Market Whipsaw, and Earnings That Tell a Different Story

The handpicked topics for this week are:

1. Data Centers Land on the Ballot: Patrick Moorhead walked through the results as reported: Democrats flipped seven Senate seats for a 54 to 44 majority and won the House 259 to 176, and economic concerns topped exit polls. In Texas, which Moorhead noted has the largest planned data center buildout of any state, Gina Hinojosa defeated Greg Abbott after campaigning on a data center moratorium and an end to developer subsidies. Gubernatorial winners in Ohio and Georgia also backed pauses. Voters in 10 states weighed 41 local data center measures, and every one with reported results went against development except in Harwood, North Dakota. Daniel Newman noted that Abbott had already put a moratorium in place months earlier and read the vote mainly as a referendum on affordability, taxes, and schools. (The Decode)
2. From Euphoria to Despair in Two Days: The hosts traced the selloff to concerns about neocloud financing, Elon Musk's prediction of "guppy-level death" for 200-plus AI firms, and a slowdown in Google's search business despite 100% growth in its data center business. Microsoft and Amazon jumped more than 6% Wednesday as investors bet on hyperscalers over neoclouds, then Thursday wiped out chip names: Moorhead cited NVIDIA down 7.5%, and CoreWeave down 12%, and Newman noted Micron down close to 20%. Newman argued the market is pricing AI as if the buildout will stop, a reading he called wrong. (The Decode)
3. Where the Data Centers Go Next: Newman expects every project slated for 2027 and 2028 to move forward, with tougher community terms and fewer tax breaks for 2029 and beyond, and with states like West Virginia, Idaho, and Oklahoma eager to compete. He shared research estimates of $3.1 trillion in 2027 capex and $4 trillion in 2028, with a gridlock downside near $2.8 trillion. Moorhead pointed to federal land, Eastern Europe, Malaysia, the UAE, Saudi Arabia, and even space as alternatives if U.S. states pull back, warning the country could give up tax revenue and construction jobs.
   (The Decode)
4. Export Controls and AI Guardrails: Moorhead expects a Democratic Congress to push for tighter controls on advanced semiconductors and AI, along with public AI hearings, guardrail legislation, and scrutiny of SpaceX and Tesla federal contracts. Both hosts expect plenty of gridlock and believe any changes will arrive more slowly than the market fears. (The Decode)
5. AMD ($AMD) Becomes a Data Center Company: Newman highlighted 97% growth and a guide toward $17 billion next quarter, driven by Instinct GPUs and EPYC CPUs, including the CPUs powering Meta's Muse agents. Moorhead noted data center now accounts for 64% of revenue, with Q4 likely the first quarter data center AI clearly outsells server CPUs. Lisa Su said wafers, substrates, and memory are secured, yet demand still tops supply. (Bulls and Bears)
6. Apple ($AAPL) Posts a Record but Faces Memory Costs: Moorhead said Apple printed a record September quarter, with memory costs set to show up in the holiday outlook and supply capping holiday upside. Newman framed Apple as a flight-to-safety play built on low capex, wide distribution, and strong product margins. (Bulls and Bears)
7. Qualcomm ($QCOM) Diversifies Beyond Handsets: Newman called it a reset year as the Apple arrangement unwinds, offset by triple-digit automotive growth and data center deals. Moorhead highlighted a reaffirmed $5 billion data center target for FY27, with $650 million pegged for next quarter and a $7 billion-plus exit, along with a new SpaceX Starlink partnership for radio and baseband infrastructure and a high-bandwidth compute architecture positioned as an alternative to HBM. (Bulls and Bears)
8. Arm ($ARM) Rides Data Center Royalties: Arm grew 27%, with licensing up 50% and royalties up 8%. Moorhead noted data center royalties doubled as accelerator head nodes move off x86, while commitments for Arm's own AGI CPU doubled from $1 billion to $2 billion ahead of first deliveries in Q4 2027. (Bulls and Bears)
9. Lattice ($LSCC) Grows Well Beyond the Acquisition: Moorhead said revenue more than doubled, mostly organically, with FPGA revenue up 89%, industrial and embedded up 70%, and margins beating guidance by two points. Lattice and AMI are positioning to become the control plane for AI racks across any accelerator. (Bulls and Bears)
10. Coherent ($COHR) Proves Its Optical Scale: Moorhead pointed to a $20 billion, five-year PhotonLink agreement, NVIDIA's $2 billion investment in March, gross margin up 184 basis points with a raised 45% long-term target, and indium phosphide output that doubled year over year. Newman argued the pending Marvell and Lumentum deal makes Coherent even more strategic as co-packaged optics demand grows. (Bulls and Bears)

Watch the full episode at sixfivemedia.com and subscribe to our YouTube channel so you never miss an episode.

Disclaimer: Six Five Media is for information and entertainment purposes only. Over the course of this video, we may discuss companies that are publicly traded, and we may reference their equity share prices. Nothing discussed during this webcast should be considered investment advice or a recommendation to buy or sell any security. We are not investment advisors, and you should not rely on this content as financial advice. Six Five Media collaborates with technology companies and industry leaders to produce research-driven interviews and multimedia programming for enterprise technology audiences.

Daniel Newman:

‍

Do you, Patrick Moorhead, on this day, November 5th, 2026, one of the worst days in market history, actually still believe you know what you're talking about? 

‍

Patrick Moorhead: 

‍

I don't know, Daniel. I'm gonna have to go review everything I said. I never quite said that I was ever a stock market expert. And in fact, typically I am the buy high and sell low guy. Currently I am the buy high and not sell and hodl at this point. So.

‍

Daniel Newman: 


This is what the retailer community would call you now a bag holder.

‍

Patrick Moorhead: 


Exactly. And the last, the last like 15 minutes of today was brutal.

‍

Daniel Newman: 


Yeah, I've never seen anything like it. And I've seen a lot, you know, it's funny, because today was a day that sort of reflects a few days in history, days that you've been through as an adult days that I went through at my just into my adulthood, but I was thinking, you know, you had dot-bomb. You had, I'm sorry, you know, you had dot-bomb 9-11, that whole weird window. You had the 2007-8 mortgage crisis. You had COVID. You had the massive deleveraging of 22-23, one of the worst sell-offs I've seen in my lifetime. And then you had Liberation Day. And it's funny, because it's getting more, the cadence is getting more regular here. Like between 2020 and 26, we've had what, like, at least four capitulations, like real capitulations. And today, I mean, the Nasdaq up close to 6%. And we know this isn't like a stock show per se, but everything that's happening in the market right now is all so closely interdependent, interconnected, right? So this is a special episode for everybody out there. What number is this again? I forget, but it's, I have it in front of me.

‍

Patrick Moorhead: 

‍

326.

‍

Daniel Newman: 

‍

No, it's not 326, but 326, we're going to do a little different. We just feel like sometimes there are moments that are bigger than a specific formulaic podcast. We normally do our decode, our flip and our bulls and bears, you know, where we go through kind of all the news of the week and then we argue something and I win. And then we talk about all the week's earnings or market moves. But this week we kind of saw like this, confluence of forces. You know, we've seen this massive AI buildup slash bubble that people have been challenging now for three or four years. We saw a national election of enormous magnitude that took place this week. And one of the things on the ballot is, of course, data centers, technology. But there was also a lot on the ballot about the economy, about affordability, about education, about social issues, a gamut of social issues, taxes, etc., etc. And there was a pretty strong referendum this week in the vote that indicated that the populace at this point has a very different idea about where they want this country to go than where it's going. And so that, of course, intertwined with a massive technological buildup that's been heavily supported at the federal level, at least at the executive branch. Working tirelessly to see the United States become the world leader in AI sort of may have been unwound this week in some ways, at least with the direction of the country. And there's a lot of other nuances in there. And again, it's not a political show. So we may delve into parts and pieces of that. But I want to kind of focus on the technology angle here, but what we saw today, you know, you could use the word capitulation. We actually saw a two-day swing from, you know, basically market euphoria to despair. Patrick was getting therapy earlier from his ChatGPT bot.

‍

Patrick Moorhead: 

‍

Yeah, we really thought yesterday that the coast was clear, the idea that data centers would be paused everywhere, that the day before we're incorporating it. And let's just be clear, where this all started was financing of neoclouds. Do you agree with that? And that is where the pain and suffering.

‍

Daniel Newman: 

‍

I think that was one of the first kind of in the unwinding that we've seen. This particular cycle of FUD or of the end of AI started with our neoclouds. Yeah. Do they have any utility? Are they going away? Do they have any money? Are they going to be able to finance? Can they actually do what they say they're going to do? And of course, Elon Musk's comments about what, 200- Guppies, yes. Guppies, level, extinction, death. Yeah, so that started it, but then you had earnings. And even in those weeks though, Pat, we saw ups and downs, like these huge volatile ups and downs. Microsoft just crushed it. Meta, you know, looked pretty good. Everyone was affirming their CapEx. And then Google, which by the way, laid out a freaking monster growth quarter for AI, 103% cloud quarter. But their ad business, search business, decelerated. And that felt like the next unwind. It was like these businesses, Pat, had like a disconnect. Their core businesses that were funding the AI started to look like they had cracks.

‍

Patrick Moorhead: 

‍

Yeah, and the Google one is super ironic because the data center had 100% growth, right? They literally doubled their size and their profitability was a sight to behold of. And then we got to election day. And here are the facts, right? On Congress, I mean, AP called the majority control of the Senate at almost midnight Eastern. Democrats flipped seven Senate seats and hold 54 to 44. I think Alaska and Mississippi might still be in called. And then in the House, 259 Democratic seats to 176 Republican, an absolute destruction of Republicans and who they had voting. And we had a lot of downvote impact, right? People coming out In my opinion, it was the anti-Trump vote, and that just carried on, right? If you came in to go against Trump and Ken Paxton, you're gonna vote D all the way down the line. And that's a lot of what we saw, and there were some nuances in what was coming out, but that was the general trend. We also saw a lot of people who voted for Trump two years ago, a vote Democrat this time as well. And it was pretty clear what the causes were. And it was on economics were the number one thing. And this kind of makes sense, right? Republicans came in for the presidency and really talked about border control, but also the economy, right, bringing the level of cost of goods down and bringing interest rates down. And, you know, the public was very annoyed by those things. And Iran conflict came up a couple times. It wasn't as big as everybody thought, but I think that was primarily related to a bad sales job by the administration. But also, if you remember, we had two, you know, whacks at this where we had bombed Iran, specifically their nuclear sites and said, hey, we have disabled them, then we go and then we attack them. And, you know, the first strike was crazy and ended up with the removal of pretty much the entire leadership and military staff. And then essentially Mission Accomplished came out, just like George W. Bush had put out there. So again, a lot of anger on what was promised and the perceived notion and a lot of reality of what was delivered and what wasn't delivered.

‍

Daniel Newman: 

‍

Yeah, I think you characterized a lot of that well. And it's interesting because clearly, as I mentioned, there was a referendum. Now, I will say this without comment. This election brought so much enthusiasm.

‍

Daniel Newman: 

‍

Yes.

‍

Daniel Newman: 

‍

That we saw 2 million more people vote for the governorship. And I think it was more than maybe even closer to 3 million more for the senator than ever in history. So it's either all the imported, like from California and that voted, but very interesting data. We actually had more votes in this election for the railroad commissioner than we had for governor in 2022. So clearly people deeply cared about what was gonna happen in Texas during this election. snark aside, whatever you wanna do with that, do with that, like I said, data without comment in this case.

‍

Patrick Moorhead: 

‍

Let's answer. Yeah, let's bridge to the tech part.

‍

Daniel Newman: 

‍

Yeah, I want to talk about this, though.

‍

Patrick Moorhead: 

‍

But it's very clear to talk about, like, first of all, who won that mattered, right? So Texas has the largest future buildout of data centers of any country in any state out there, right? And Democrat Gina Hinojosa beat Abbott after promising a data center moratorium and an end to developer subsidies and tax breaks. Ohio's Amy Acton backed a conditional moratorium and Georgia's Keisha Lance Bottoms proposed a temporary statewide pause and then local ballots right voters in 10 states considered 41 local center measures, everyone with reported results went against the development, except for Harwood, North Dakota. And then on Wednesday, Microsoft, Amazon, funny, rose more than 6% because this was the anti-trade to the, oh, they can do it better than the neoclouds, and CoreWeave fell 10% And then Thursday, which we're in right now, obviously, you know, a complete wipeout, right? NVIDIA down 7.5%, AMD down 15 points, Core down 12.

‍

Daniel Newman: 

‍

Micron down close to 20. But that's also many of these names down already 40, 50, 60% from their highs. Now, I want to put a little context pattern on what you just shared, though, is First of all, like just generally speaking, data centers while very well supported at the federal level from the executive branch have been very mixed in terms of what level of support they had from both Republicans and Democrats. Now in Texas, current governor Abbott had already put a moratorium in place several months ago. So that's first of all, that's noteworthy. Another thing of noteworthiness is why did the market have that sort of euphoric and then despair? And I think the euphoric moment came because there was this underlying belief that some type of action that would come from these newly elected Democratic governors to slow down data centers was going to mean that suddenly Google, Amazon, Microsoft, and everyone else was going to stop building data centers and stop working on AI. It's the only thing that made sense because they were going up parabolically yesterday, kind of like indicative that they were going to return to high free cash flow, spending less money on AI build out. And at the same time, chips were dropping and data centers were selling off as if they were going to be left abandoned, like unbuilt housing communities in 2007 and eight. And let's be very clear, like Hinojosa and Abbott and all these governors in different states, this is something I think is really, really important to notate. Texas is not yet, and I won't say it will never be, it is not yet a deeply blue state. This election, you absolutely have to accept it for what it is. This was an absolute referendum on, you know, The affordability being poor, prices are high, taxes are high. The market has not performed particularly well for most of these people. Schools have suffered. The voucher program is disliked deeply. There was a lot of things that Abbott just didn't execute. And it had been, what, almost two decades in power. And I think there was a sort of a, hey, you know what, we may or may not agree with the Democrats' views. We certainly, I think, in Texas do not agree outside of some larger metropolitans with the DSA, the social side of this. But what we do believe is we want affordability. We want good deals in our communities with data centers are brought in. We want infrastructure to be built out. We want our taxes to be lowered. The property tax here is a big story because if you don't pay state income, but the property taxes are absorbent here in the state of Texas. All that is happening. The reason I'm just pointing to this is, Data centers are the GDP of the state of Texas, and they actually are the GDP of the United States. Data center and AI is actually 100%. This is debatable, but I believe it's shown up in a number of economists' models. 100% of the GDP in the United States basically ties to AI right now. Okay, the growth, GDP growth. Texas, 25% of those data centers, I think you called that out earlier, If for any reason, a new governor just comes in, shuts that down, they're going to shut down jobs that are on people, on sites, doing jobs, building stuff out. The actual short run here, Pat, is every project for 27 and 28 is actually going to go forward. That's right. And that's an important note, though, for people, because we're selling the market like it's just going to stop, like that NVIDIA is going to have a guide of 10% of their, you know, originally forecasted to the next quarter, that, you know, all that Microsoft's going to just cease building data centers, Stargate is going to get canceled. I mean, that is how the market is reacting right now, and that's not real. What's going to happen is the deals for 29 and beyond, and I'd like your take on this, Hinojosa, the governors in Ohio, South Carolina, and Georgia are going to go back on any data centers that are being built, and they're going to demand higher thresholds, more investment in communities, likely less tax breaks, but these data centers are not going to not get built. And there's one other thing I'd like to get your take on this, and then this, Pat, is The other thing is that you have states now raising their hands the way Texas used to, West Virginia, Idaho, Oklahoma, that are kind of saying, we're open for business. And obviously 27, 28 is in the books, but some of this 29 and beyond, it may be less in Texas, less in Georgia, more elsewhere. What do you think on that?

‍

Patrick Moorhead: 

‍

Yeah, I mean, we've already seen some of the states reaching out of who's open for business. And the other side of it is federal land isn't actually governed by the states. And there's two data centers going up on federal land and plan right now. And that is a potential opening for the future. But the one that I found Even more interesting, Daniel, was this notion of different countries. And I talked to a couple of CEOs this week and had discussed this. And then I even saw Rene Haas talk about on CNBC. And don't be confused. Arm's Rene Haas is not only the CEO of Arm, but he is the number two at SoftBank, and who is underwriting a lot of the data center activity already. So he knows, and he has got a better global view of this than many. He's not just the chip IP company. Essentially, if the US doesn't want this, then other countries will take this with open arms. And you're like, well, Pat, where are these? And why weren't they there before? So Eastern Europe is a great place, because not only do they have energy, is they have a desire to add to their economy. And they want to be seen as a tech player. And whether that's Poland, and I know Kazakhstan was thrown in there as well. But also, let's not forget about Malaysia, right? Malaysia that has come up in a lot of times about, oh, I think Meta made a comment about Malaysia, or maybe it was just pulled out of a transcript somewhere, but their net is there are other places and other countries. And by the way, I'll put the UAE and KSA in there. And yes, they have this potential risk of military conflicts. And the AWS data centers were attacked. But there are other places in addition, and it's funny, Elon even talked about, you know, space data centers that I know there's a lot of debate and consternation about that. There's a startup, I've forgotten their name, that already does this. But then you have Elon, who is very dead set, and investments from Nvidia, and quite frankly, We already have many data centers in space with SpaceX. So if the US and certain states don't want to do it, we will take it to other place. The final answer to my own question is, well, why were so many of these done in the US? And they were to pacify and please the current administration. And Trump made a commitment is winning the AI race, where he had talked about, he signed two EOs. I was at the event on getting ourselves out of the way to be able to get much easier licensing. So, yeah. I think we have fallen in love with ourselves. And unfortunately, as a country, we might be passing up tax opportunities and construction jobs. And by the way, taxes on every piece of material that get put into a US data center, we would be potentially missing. And it's so pathetic, too, that every data center story does not have a sad ending. I mean, Microsoft, in one of their largest data centers, essentially the taxes that came out of that built an entire school, built an entire hospital with that tax money, improved roads. So, you know, if we want to let this slip, c'est la vie and FAFO, if I can.

‍

Daniel Newman: 

‍

Well, Pat, I mean, the reindustrialization of the US has been basically enabled by this AI revolution. It's not even debatable. I mean, we've brought, this is the largest boon. And people say, oh, they're temporary jobs. What construction isn't? What infrastructure isn't? When we did deals in the Obama administration that focused on adding roads and, you know, we created tons and tons of jobs, we raised, you know, We did tons of infrastructure deals and roads are temporary. You build a road, you go build another road. Data centers are the same. We're going to need to keep adding more capacity and more data centers. And these jobs pay very well. This is very lucrative work in these communities. And to your point, the money that comes in, and this is where maybe everyone could do better. This is where maybe the new Hinojosa or other administrations across other states or maybe the new federal leadership could do better is, yes, these companies will certainly be very profitable from building these data centers. Can you negotiate it out, more dollars to build schools, to improve roads, to bring hotels, hospitals, retail, and other amenities into these communities that make them stronger in the long run? Can you beautify these communities around these data centers? I mean, sure. I think there are dollars to be extracted, extrapolated, extorted, whichever word you want to use, on the behalf of getting these things through. But I like to use the term like it's all in the margins here. I think left or right, keeping this moving is actually prudent, if not existential for the growth of the United States. I do think your point about the fungibility, though, is very real. I mean, the Kazakhstan thing was timely. Malaysia, I think, came out today.

‍

Patrick Moorhead: 

‍

Regions in South America, across northern Africa, across Saskatchewan, I mean, which, I believe, I hope I get this right, is the Canadian province with all of the all of the oil and and it's cold. Right? Could be a perfect place to to do something like this if the Canadian administration doesn't get in the way. And by the way, please don't get me wrong. I want this to happen in the United States. But overall, the most important thing to me is that it happens.

‍

Daniel Newman: 

There's one other thing to note though, too. I mean, I think you had some of these conversations. You mentioned your CEO combos. The data centers go on and get built in other places. And the chips will be absorbed into other infrastructure that's going to be built by companies all over the world and for technology all over the world. You know, we're so short memory right now. I mean, Micron, I think it's being priced. It's almost to the point where it could buy all its shares back over like two years. That's how like mispriced it is based on its expected earnings and cash pile. It could literally buy itself back from the public. I mean, it's insane. I think Nvidia following today's route is like down to like a 10 times forward. I mean, I just think that, you know, the market is treating this like like it's dead. And I think it's really wrong. And by the way, our ETR equities research team published today, sent out to our financial service clients, and I'll share this with our community here, reiterated the CapEx for 27-28, changed a few of the scenarios, but basically 3.1 trillion for 27, 4 trillion for 28 estimates. What we did is we noted that if the note from Rolf Bulk that runs that research basically was that it would be political suicide. And this is what I was trying to say about like, if they want to stay blue, because like for like a Texas, for instance, or Ohio, they want to actually elect the next president. What they need is two really strong years, strong economic years, strong growth years. They need to basically be able to say, look how much better things got in 2026. Because states like Texas and Ohio will revert to the mean and will go back if they feel that affordability and taxes and the economy were not the focus. If it becomes a, you know, a more of a focus on social issues and stuff like that. I don't think Texas, and God forbid someone starts to try taking your guns, like that will, Texas will run back in a hurry. So they have a chance right now to deliver and I think allowing this to play out. So basically we have like a worst case scenario is if everything hit gridlock, you could see the 4 trillion drop to closer to 2.8 trillion in 28. And I guess that's the downside scenario that exists. That's still massive and that's still probably full of all kinds of constraints on chips. Yeah.

‍

Patrick Moorhead: 

‍

One thing we haven't talked about yet, Daniel, is the potential export controls. I think we can, there is an expectation that a Democratic Congress will likely push for tighter controls on advanced semiconductors and AI. If you remember, I know Biden's executive branch, but the whole diffusion rules, and kind of where that got us. And then, you know, Trump administration used export controls, I think, as a wedge on tariffs. But there is a belief that there will be a lot of that. Heck, we even saw Ted Lieu come out and say there's going to be oversight in hearings, public AI hearings, pass guardrail legislation, and consider redistricting new data center construction. So the incoming House Oversight Chair posted, quote unquote, lawyer up. And Democrats have discussed probing SpaceX and Tesla federal contracts. So there we go. I mean, those are some downstream impacts that- Do you buy it or political theater?

‍

Daniel Newman: 

‍

What do you think? Sorry? Do you buy it or is it political theater?

‍

Patrick Moorhead: 

‍

I think we will have hearings and it can be both.

‍

Daniel Newman: 

‍

Do you think, oh yeah, it's true. But do you think the Bernie Sanders, 20 years in jail for using super intelligence, like do you think that stuff lands? Do you think we have a strong enough anti-AI constituency that that stuff's gonna garner? No.

‍

Patrick Moorhead: 

‍

No. And I don't understand the timing, because the timing is typically around fundraising. And I feel like that time is over. And what does it benefit you to have you know, fundraising, uh, I get it two years from now when, when you're going to need the big money to go after the presidency. And, you know, there's a lot of talk about AOC now has a really good run. I hear the ability to run for president. Um, but I just don't understand. It seems like a lot of, you know, old man get off my lawn, uh, type of, uh, uh, type of talk, but, um, yeah. I, I, I, every time Daniel, I think that cooler minds will prevail and common sense will prevail. I am universally disappointed. And those times I get more of those every single year related to, uh, the political, uh, mess.

‍

Daniel Newman: 

‍

Yeah. I mean, I think in the end, um, We'll wind this part down and we'll head in and talk a little bit. There was a bunch of earnings this week. Again, I just don't know if anyone cares. We'll care again next week. We'll be back. We're going to talk about those. You know, I think if the change is alarming to you, I think we found that these kinds of regime changes certainly have impacts, but I think they come slower and they take longer to be felt. I don't think that like a switch was flipped. And I mean, we still have several months between now and when these people are even sworn into office. As President Trump said, I'm still your president. You may hate that. You may like that. But right now, we'll have a divided set of chambers for at least the next couple of years. So you can probably expect a lot of gridlock. In Texas, the state legislation here is going to be mixed. It's going to be a Democrat House with a Republican Senate. Very odd that it worked out that way, because it was so convincing, but it is, with a Democrat governor. And so she can't unilaterally, just like Trump can't unilaterally do a lot of things. It's kind of one of those things where I actually think there's going to be a forced function for some aisle crossing to take place for things to pass. And I think the stuff that's going to happen has to be focused on affordability, and on schools, and on things that are not these fringe issues But that, by the way, you know, the credit to the to the Democrats here in Texas was the right. The Republicans ran on so much fringe bullshit. Yeah, I really wish they'd ran on some real issues, because I really don't know. But I mean, in the end, it's like the entire thing was about gender, about like, just weird stuff that like, I think some people care about, but I just don't think that's what when people are paying four or $5 for gas and seven for diesel, and their taxes are higher than they've ever been. I mean, I think most people are, they're focused on their wallet, their checkbook and their families.

‍

Patrick Moorhead: 

‍

Yeah, listen, the exit polls clearly show that there's no ifs, ands, or buts about that.

‍

Daniel Newman: 

‍

Yeah. So, Pat, that was fun. It was different. You'll have to tell us if you liked listening to Pat and I talk about one thing for 30 minutes, or if you would prefer that we move around a bit more because we droned on too much. I'm letting you reflect. I'm letting everybody reflect on that now. All right.

‍

Patrick Moorhead: 

‍

Let's close out the uncomfortable pause, Daniel. I appreciated that.

‍

Daniel Newman: 

‍

Yeah, I know. I'm here for you. I'm media trained, in case you didn't know that.

‍

Patrick Moorhead: 

‍

Somehow I think you can do the training.

‍

Daniel Newman:

‍

 There's a lot of earnings this week. We'll do a little bit more rapid fire, but I feel like big ones. Let's just march down the list. Let's hit the bulls and bears. Cool, man. I'm going to call my own number first, but I'm going to totally share and give you some limelight, because I know this is a big topic, Pat. But just for everybody out there, we're going to hit AMD, we're going to hit Apple, we're going to hit Qualcomm, we're going to hit Lattice, we're going to hit Arm, we're going to hit Coherence. There's a few others we may or may not get to from there, but that's a lot. Like that was, I actually ran out of breath just going those, but Pat, like AMD brought the heat. Lisa Su notoriously under promise over deliver. She under promised last quarter over delivered this quarter. And I think under promised again, which she's going to deliver next quarter. Now, this will be some of that sort of underlying proof of whether or not everything stopped, which it hasn't, because she's going to see her sales swell. And by the way, it's something for everybody out there to listen to is listen to see if any companies actually adjust their forecasts. You know, public companies, if there is a material change in their expectation, can't wait till the next earnings. So all you think the data center thing is dead, all these companies are gonna have to come out and confirm if their misses are gonna be more than nominal coming into next quarter, sidebar. But like, this was a great quarter, great on the top, great on the bottom, a 97% growth, estimating next quarter, Pat, 17 billion next quarter. I think they're getting to the point now where they're doing per quarter what they were doing for a year, not that long ago. They of course have these two businesses that are absolutely going parabolic. One that everybody expected, which was their Instinct GPU business. One that not everybody until agents really popped, started to fully expect, but is the EPYC CPU. And she made a big note about the fact that all that meta growth of their Muse agents, all that's running on EPYC. Um, you know, I kind of left my interpretation, Pat was, you know, the ramp is going to be hard. You saw a little bit of that showing up, like, can they keep margins and the integrity of their margins high as they're ramping up Helios? But like, I think, you know, She's conservatively guiding to growth, and I think she's got the contracts in hand, and I think she's got a lot of reason to feel confident that she can keep going. And for the investors out there, company's a whole lot cheaper than it was a few weeks ago, but nothing about its actual fundamentals has changed.

‍

Patrick Moorhead: 

‍

It's a good place to end. Yeah, some of my biggest takeaways, I mean, AMD is now a data center company with a PC business, which is just so ironic because AMD was a PC business with a kind of a bolt-on data center business. Data center is 64% of Q4 revenue, and that's pretty crazy. And it was interesting, Q4 is going to be the crossover for data center AI. The first quarter data center AI outsells server CPUs by a wide margin. AMD doesn't split apart CPUs and GPUs in the data center, but it looks like this is the crossover. And this is where stuff gets real. And people I know are I don't know, LARPing about AMD not having, you know, A, not executing well, and B, always being behind. It's funny, people forget where AMD came from. AMD was almost bankrupt, three months of cash left. My strategy was don't, by the way, I ran it, I ran it up, ran it down, ran it back up, and then down and almost death. You brought that up, I have to defend AMD of old. Two mistakes. One, a really bad microarchitecture for a CPU called Bulldozer. And that drained cash away from almost everything. And secondly, paying too much for ATI. We got sold that it was top of the stack, and it was N minus two. And the value of the company should have been about half. So yeah, we had to sell assets, we went from top of the stack to literally entry level maker and AMD to their credit, locked in on Zen architecture, only had money for one CPU, and they didn't have enough money to do multiple monolithic designs, and they made chiplets, high performance and low power, which is part of the reasons they are where they are today. But I digress. The other thing I liked was Lisa came right out and said, AMD's wafers, substrates, and memory are secured, but demand still tops supply. And this was a Q4 statement, but she did give some cookie crumbs that says first half 27 isn't a done deal, but they felt pretty comfortable of where it's gonna go. So the company guided 3.6 billion above its own record quarter, and now the PC business is the footnote.

‍

Daniel Newman: 

‍

Well, there you go. Speaking of footnotes, probably the world's favorite company right now, CapEx Lite, innovation free. Apple. Just kidding. They have a foldable device now, Pat. So I mean, they're pretty hot and on the front edge. But it is like the most valuable company in the world now. And they had a pretty good quarter.

‍

Patrick Moorhead: 

‍

Yeah. Yeah. I mean, they printed a record and got it below the bar. But the holiday quarter is where that memory bill is going to show up, right? The September quarter was clean, even after the tariff refunds. But again, the guide is where the memory costs are going to show up. And unfortunately for Apple, their hubris and negotiating tactics of yore are going to come and bite them in the booty. Apple's done a good job buttressing this from new ways to buy with leases that Apple is just so good at this. I think the year before they basically had an all-in subscription to kind of take you over the line if they raise prices. And they did raise prices as well. So as for monetizing Siri, I think they're still doing experimentation. But if you remember WWDC, they're going to give a base level and then have a tier that where you actually have to pay. But Ternus is pricing Siri for, you know, pricing for share. And, you know, at some point, we'll we'll monetize monetize Siri. So supply is really capping the holiday upside. And they were very, they were very clear on that. And I'm not just being confused by the iPhone duo, but really the whole cadre of phones up and down the chain.

‍

Daniel Newman: 

‍

Yeah, I feel like This is a flight to safety play, low cap X, high distribution, strong product margins. They're clearly fighting the memory battle still, which is probably a weird thing for them. They've never had to fight that kind of battle before. Um, but you know, they're doing well and I think, you know, they will continue to be kind of seen as a safer play because that's just the way they're constructed. Yeah. So, you know, again, it was a, you know, you got the holiday season coming. Um, you've got the new products. There was very little reflection of the new products in this quarter numbers. And so, you know, we'll see how those new devices are selling. Did you get one of those yet? The duos?

‍

Patrick Moorhead: 

‍

No, I, I I'm finding more joy in experimenting with DGX sparks and RTX sparks, laptops, a lot more, a lot more fun. Well, nobody's ever said you're fun, Pat. I mean, nobody has said I'm fun, particularly since I pulled back on staying out till 2 a.m. in the morning.

‍

Daniel Newman: 

‍

Used to be more fun.

‍

Patrick Moorhead: 

‍

You used to be fun too, bro. So now it's really about how early can I get in bed? And how much protein can I slug back? And how much pre can I ingest before the workout? And how do I not throw up after my workout?

‍

Daniel Newman: 

‍

That's good stuff. I mean, we were going to do an episode on me shoulder pressing for plates today. But then we have to broadcast it on X, like, I mean, I'm just I am feeling with I can't seem to get anything right in the market. So I gotta get a different kind of pump right now.

‍

Patrick Moorhead: 

‍

No, you just adopted the Pat Moorhead investment strategy, it looks like.

‍

Daniel Newman: 

‍

Well, the thing is, is like I want people to be afraid to tell me I'm an idiot because they know I might eat them for breakfast. All right, Pat Qualcomm. This was like a tricky one because like they weren't expected to grow. This was actually a down quarter, down year. They were unwinding their Apple arrangement. They are diversifying business. Their core businesses are in areas that aren't really hot right now. They're fighting the same input supply impact. On top of handsets not being as cool as they used to be, you're also paying more to build them. Sales are down in that area. But on the other side, this is a company that's having a bit of a coming out party for all of its other businesses. And I think that's why I've been way more optimistic, especially as a company that trades so cheaply, automotive business getting your triple digits, it's non handset businesses are surpassing its other, the rest of its businesses by next year, that'll be a handset will be a small fraction of the overall business of the company with the $5 billion plus of data center compute chips, and they've already got deals inked with Amazon. I believe ByteDance, but I think there's others in tow as well. And they're doing some cool things outside of that with networking, a very diverse set of capability across network optics, compute CPU and custom AI chips. And so that's gonna rise to about 15 billion over the next couple of years. But really, like I said, this was a reset year, getting the old out, getting through some of the historic challenges and getting the new in. They've shown the ability with some of these new businesses to ramp. I think this AI one will be the fastest ramp. And then on top of it, right on their earnings day ahead of their release, they were able to announce, which I guess it got less popular with the election results later in the day, but they did announce a partnership with Starlink, which, you know, basically they become the, you know, layer one platform and they become the hardware and infrastructure for 5G, 6G, AI, native 6G, AI. And this is like radio and baseband infrastructure, not handset chips. And this is really interesting because you and I have been somewhat outspoken, like, can SpaceX really pull this off? This definitely adds some fuel to the fire. This is a partner, this is a company that's done this before. This was sort of deep inside of its genetics. It's something they've been sort of away from for a bit now, and now they're back. And they're aligned with a company that's making very interesting overtures into what we would say a heavily incumbent based and very established market. And this could be a great diversification into their physical AI Dragonwing sort of broader portfolio, which we've been optimistic is a nice growth option for the company, Pat. So, you know, they had that and, you know, I'll give you some oxygen here. I know they made some other announcements about HPC, et cetera, that maybe you can read it on.

‍

Patrick Moorhead: 

‍

Can I go now? I almost fell asleep. Is it that long? Racing on the show right here. Is that any good? I'm just kidding. I'm just kidding. No, listen, I mean, the car, the tower and the data center carry the story, right? If you want to evaluate the company on your spreadsheets with no insights, then you're not going to be as impressed. What I see is a diversification story that gets more real by the day and gets bigger by the day. And it's funny, the thing I got most excited about was the $650 million that they pegged for the next quarter for data set, right? That says this stuff is real, okay? And it's one thing to, you know, say you're going to do something. And it's another thing to load in the revenue. And what this did is, you know, they reaffirmed the $5 billion for FY27. And also, you know, showed how it's going to rise every quarter to a $7 billion plus exit. So really impressive, takes it from theory and thesis. And remember, guys, This is supposedly a smartphone company and they still are. OK, which, by the way, they said there was going to be a trough and they're out of it at this point and was really enthused to see that China is at a bottom and Qualcomm dominates in China premium and premium as a whole. And they said that they were moving their mix in Samsung from 50% to over 70%. So positive stuff there. The company did do another disclosure for data center that I thought was interesting, HBC, which is called high bandwidth compute. It's an alternative to HBM. And what it does is it, layman's terms, it actually fuses the memory super close to the compute, and the bandwidth of it is absolutely extraordinary. And they also came out with the power. And again, this is for next generation HPC-based accelerators, but also for processors as well. And by the way, Nobody will tell me this on the record, but when I do talk to the big CPU and GPU and accelerator companies, they all acknowledge that this could be the future. And you had to know that Qualcomm was going to come up with something that had to be architecturally different and something that was more in their wheelhouse. But listen, a lot of execution to go, but I do like that they have these major customers showing up and they are driving revenue off that.

‍

Daniel Newman: 

‍

Yeah, I think some of their like the optical skills they have in their back pocket, you know, you notice noted that Lumentum Marvell deal is a big one this week, but obviously fell flat on its face in the public markets. I think just out of whether it can get done. But I think the big bet is that Optical Interconnect is going to be huge and a company like Qualcomm could partner with a company like Lumentum, take some of its IP and skills and actually build some really interesting

‍

Patrick Moorhead: 

‍

Uh, capabilities in-house and could partner with some. It's just a, these are things that people don't think about balkan for, but they have a lot of that kind of IP lying around. Well, it's cool when you don't have a base to protect of what you can do architecturally and then partnering, right? It's actually pretty fun. It sure is. I like to partner with smart people. I like to partner with you. How many like, we have like seven signed documents, don't we?

‍

Daniel Newman: 

‍

Seven children, and they're all have rent due. None of them have jobs. I know it's terrible. It's terrible. All right. Arm, Arm fifth, I think straight quarter with 20 plus percent growth. Up 27%, license roaring 50%, tons of strength in his data center business. I love, love, love that they have this 60% of data center CPU claim. That's wild. And that also just goes to show the diffusion and then the adoption of in-house silicon is very real. So as we've continued to have these debates about merchant silicon versus in-house and homegrown silicon programs, and the kind of cool thing for Arm is that they kind of get it all no matter what. They're winning with the in-house programs. They're winning with NVIDIA now. They're just tied in at so many different levels. And of course, now they're building their own chips. And they guided, I think now that they see clarity to that $2 billion incremental growth, in their business through their chip manufacturing in their AGI CPU. You know, it was clean, it was tidy. This company's always been on the higher end of the valuation.

‍

Patrick Moorhead: 

‍

Is there any company with a higher forward PE?

‍

Daniel Newman: 

‍

No, they're getting the respect for the fact that they are attached to so many parts of the business. And they do have a kind of an interesting play in edge AI robotics out there as well. And it's another company that could see a lot of benefits from that market starting to mature and materialize back. But no, clean, tidy quarter. I'll pause there.

‍

Patrick Moorhead: 

‍

Yeah, I mean, listen, I noticed I skipped one, by the way, so we'll go back to that one. So, yeah, I mean, on the basics, Arm's license grew 50% and royalties grew 8%. And the market priced the royalties before the big cataclysmic Nasdaq decline. Standouts for me, I mean data center royalties. Structural shift here, right? Data center royalties doubled and the company said they expect a triple digit growth for years as accelerator head nodes leave x86, and they cited Vera Rubin and Google pairing TPUv8 with Axion. But think about that, you make a claim about that is massive. By the way, doing this with royalties at the same time that you are selling your own chips, and in this case, AGI CPU, It's supply gated and mostly fiscal year 28. So pretty amazing that they could get out with that. And by the way, first deliveries are fourth quarter 27, OCI instances in 27. And yeah, they gone from a billion commit to 2 billion So it is amazing how competitive CPUs are, and I'm glad to see it.

‍

Daniel Newman: 

‍

Absolutely. Bring it on. Bring it on. $40 to $200 billion market. Clearly, the CPU business is in good graces. And there's some demand still. So don't cancel all your don't don't sell everything yet. Or maybe you feel you need to just don't jump. Don't jump. All right, Pat, another company that just crushed it out of the park. FPGA kind of a company that's been quietly probably performing in that category better than other lattice semiconductors.

‍

Patrick Moorhead: 

‍

Yeah, a lot of still more than doubled revenue. And the acquisition was a was a tiny part of it. So a ton of organic growth here. So, you know, FPGA alone was up 89%. Right. And Sorry, I said huge. No, no, it was huge. And the momentum is carrying into Q4, right? FPGA grew 25% sequentially, but Q4 got it up 15% sequentially, and industrial and embedded up 70% year on year. And the margins absorbed the acquisition a lot better than they had guided. A, was that Lattice sandbagging, or did they just sell a lot more higher margin goods than they thought? But no, they beat their margin by two points, and that is really hard to do. Strategically, if you look at where Lattice with AMI's capability is, they really want to be the AI racks control plane. So low power FPGAs, you've got AMI's firmware, cover control, security and management inside AI infra racks, rather than the accelerator socket. And what that does is it gives everybody flexibility on regardless of what accelerator you use, you can have a single control plane. And that just simplifies everything and lowers your actual total cost of operating. So hats off, folks.

‍

Daniel Newman: 

‍

Absolutely. All right, I will add one quick thing to that one. And that will just be that the industrial business has interestingly picked up. You know, they noted that PMI is up, manufacturing is up. Some of this is the re-industrializing and everything that's going on, but they have a data center and compute business, and then they have an industrial FPGA business. And that thing grew like 35%. And that part of the market had been absolute dog shit for quite some time. So they're firing on all cylinders, and the performance is showing. Pat, last topic, laser beams. Coherent, which you wrote a great LinkedIn post.

‍

Patrick Moorhead: 

‍

Are you coherent on Coherent?

‍

Daniel Newman: 

‍

A LinkedIn post on Coherent.

‍

Patrick Moorhead: 

‍

Flushing it. I got at least like 30 people read it.

‍

Daniel Newman: 

‍

Come on, it wasn't that many. So how did Coherent do? And is Jim and the team meeting expectations over there?

‍

Patrick Moorhead: 

‍

Yeah, I mean, Coherent proved it can make the lasers. And I think it's, you know, they just need to prove a little bit more how it can make the cash. So, Photon Link is the number that's not in the release that strategically makes a lot of sense, right? Some people call it scale in, right? Which is essentially a scale up fabric. And there was a $20 billion five-year agreement in there. And don't forget, NVIDIA committed $2 billion investment in Coherent in March. And the margin is moving for all the right reasons, right? Grew 184 basis points year over year. And CFO said it was pricing and costs reduced and raised the long-term target to a quarter at 45% gross margins. Let's talk about Indium Phosphide, right? Output doubled year over year. By the way, that's something that people seem to get no satisfaction over. The company can never do good enough on Indium Phosphide, but it seems to be growing quite handily. And then finally, co-packaged optics, bigger than the laser, the fiber, the connectors and the optics around the laser are as large as the opportunity of the laser itself, and PhotonLink sells all of it. So good stuff, good job, Jim and company.

‍

Daniel Newman: 

‍

Yeah, I'm trying to think if there's much to add there. I mean, I just think one of the really important things to point out is Coherent, I think, got a jolt. This Marvell thing, like very good for Marvell, very good for Lumentum. They can build really great stuff together if they can get this deal done, super complicated. But like that just made Coherent sort of a really an N of one. I mean, who else are these companies going to turn to for these optical solutions? They are going to be like the absolute narcos of this industry. Because obviously, some of the people that Lumentum are supplying are going to not be able to get it. So I just think it made coherent even more interesting. Um, and I think, you know, clearly the indication that we got from the Marvella momentum deal, despite the market not loving it is this optical stuff and these co-packaged optics specifically are going to be really important in the next few years. Like they're going to become really, really, really important. And I think that like people are not understanding how difficult this stuff is going to be and how, um, how limited the supply is going to be. And so I think that's why, you know, NVIDIA of course wrote a big check to both the companies. I think that's why this deal happened. And I think that's why despite, or, you know, in addition to this current quarter being good, I think that's what makes Coherent really a really positive long-term outlook. All right. I think we hit them all. Did we? We did. What'd you think, Pat? What'd you think of that format? Did you like talking like about one thing for a while or was that?

‍

Patrick Moorhead: 

‍

I did. I want to let you know I didn't prep for it. I mean, and normally I prep for these things. So yeah, I may have been fast and loose with a few of my numbers. It really just came off the top of my head. I mean, of course I did one AI search on it that took 27 minutes to complete to make sure that I had some fact checks, but no, I like it. In fact, we should think about it for the future.

‍

Daniel Newman: 

‍

I love it, and we appreciate all you. Thanks so much for tuning in. Another week in the books, episode 326, is that right?

‍

Patrick Moorhead: 

‍

Yep. 

‍

Daniel Newman: 

‍

326. The election data centers are dead for the 17th edition. Thanks so much for being part of our community. See you all later.

‍