SK Hynix Q3 2026 results show record operating profit driven by HBM4 demand, with capacity sold out through 2029.
Summary
The hosts analyze SK Hynix's Q3 2026 earnings, noting record revenue of over 100 trillion won and operating profit of 83 trillion won, while attributing a drop in net income to one-time equity fluctuations from their Kioxia stake. They highlight that HBM4 demand has sold out SK Hynix's capacity through 2029, with long-term agreements covering over 50% of production through 2028. The discussion contrasts SK Hynix and Samsung's dual-track technology strategy (4F² and 3D DRAM) against Micron's sequential approach. Finally, the speakers identify semiconductor equipment makers like Lam Research and ASML as durable bottlenecks due to their integral role in manufacturing process design.
Episode sections
- 0:01 – 3:15
Earnings Season and Industry Cycle
Hosts discuss the ongoing earnings season and SK Hynix's reporting after Micron, introducing a circular flow model of the semiconductor industry where AI benefits are recycled into design.
- 3:15 – 4:48
SK Hynix Q3 2026 Financials
SK Hynix reported record revenue of over 100 trillion won and operating profit of 83 trillion won, while net income fell 58% QoQ due to one-time equity gains/losses from their Kioxia stake.
- 4:48 – 9:04
Revenue Breakdown and HBM Demand
DRAM accounts for 75% of sales, with HBM4 sold out through 2029 and long-term agreements covering capacity through 2027 and 2028.
- 9:04 – 15:06
Capacity Constraints and Technology Roadmaps
SK Hynix and Samsung are developing both 4F² DRAM and 3D DRAM, whereas Micron plans to stick with 6F² before transitioning to 3D DRAM.
- 15:06 – 20:36
Fab Expansion and Equipment Bottlenecks
New fabs will not come online until 2027-2028, with interim retrofits occurring; equipment makers like Lam Research and ASML are identified as durable bottlenecks due to their role in process design.
Claims worth verifying
SK Hynix reported Q3 2026 revenue of over 100 trillion won, up 32% QoQ and 328% YoY.
Unidentified speaker ·3:20
How to verify: Check SK Hynix Q3 2026 earnings release or financial statements.
SK Hynix's Q3 2026 operating profit was 83 trillion won, up 38% QoQ and 632% YoY.
Unidentified speaker ·3:34
How to verify: Check SK Hynix Q3 2026 earnings release or financial statements.
SK Hynix's Q3 2026 net income was down 58% QoQ due to one-time equity gains/losses from its stake in Kioxia.
Unidentified speaker ·3:42
How to verify: Check SK Hynix Q3 2026 earnings release or financial statements, specifically the equity method investment section.
DRAM accounted for 75% of SK Hynix's Q3 2026 sales.
Unidentified speaker ·4:59
How to verify: Check SK Hynix Q3 2026 earnings release or financial statements for revenue breakdown.
NAND Flash accounted for 24% of SK Hynix's Q3 2026 sales.
Unidentified speaker ·5:32
How to verify: Check SK Hynix Q3 2026 earnings release or financial statements for revenue breakdown.
SK Hynix is sold out of HBM4 capacity through 2029.
Unidentified speaker ·6:40
How to verify: Check SK Hynix Q3 2026 earnings call transcript or press release.
Over 50% of SK Hynix's long-term agreements cover capacity through 2027 and 2028.
Unidentified speaker ·7:05
How to verify: Check SK Hynix Q3 2026 earnings call transcript or press release.
Micron plans to stick with 6F² technology before transitioning to 3D DRAM.
Unidentified speaker ·11:51
How to verify: Check Micron Q3 2026 earnings call transcript or technical presentations.
SK Hynix's last fab is slated to be finished by 2033.
Unidentified speaker ·15:28
How to verify: Check SK Hynix capital expenditure plans or investor presentations.
ASML reported better-than-expected Q3 2026 results due to upgrades to existing installed machinery.
Unidentified speaker ·16:45
How to verify: Check ASML Q3 2026 earnings release.
Do your own research
- Verify SK Hynix's Q3 2026 financials (revenue, operating profit, net income) and the specific impact of Kioxia equity adjustments in their official earnings release.
- Check SK Hynix's investor presentations or press releases for details on HBM4 sales volumes and the specific terms of long-term agreements (LTAs) covering capacity through 2028.
- Review SK Hynix, Samsung, and Micron's technology roadmaps to confirm the status of 4F², 6F², and 3D DRAM development timelines.
- Analyze the backlog and revenue visibility for Lam Research, Applied Materials, and ASML to assess the durability of the equipment bottleneck claim.
- Investigate the timeline for new fab construction projects by SK Hynix and Micron to validate the 2027-2028 capacity addition estimates.
- Compare SK Hynix's operating margin trends against peers to assess the sustainability of pricing power in the memory oligopoly.
Full transcriptexpand
- 0:00
- Unidentified speaker: All right, everybody.
- 0:01
- Unidentified speaker: Earnings season rolls on, and we have our second big player in the memory market, SK Heinex.
- 0:08
- Unidentified speaker: Of course, Micron reported early.
- 0:10
- Unidentified speaker: They always report early, but SK Samsung are up next.
- 0:14
- Unidentified speaker: You're looking at the semiconductor industry flow, and this is the updated version.
- 0:20
- Unidentified speaker: It's been retooled a number of times.
- 0:22
- Unidentified speaker: This one is very visually pleasing, I would say.
- 0:25
- Unidentified speaker: But we can take a look at one of the first iterations of this.
- 0:29
- Unidentified speaker: And it's always been that.
- 0:30
- Unidentified speaker: It's always been a flow.
- 0:31
- Unidentified speaker: It's always been in a circular format.
- 0:34
- Unidentified speaker: One thing flows into the next, and then it restarts the process all over again.
- 0:38
- Unidentified speaker: And there was a comment made by SK Heinex's CFO that I think illustrates why we have said this has been the case all along.
- 0:47
- Unidentified speaker: It is a cycle.
- 0:49
- Unidentified speaker: Yeah, SK Heinex and really the entire industry, semiconductor industry, electronics manufacturing, a lot of your favorite podcasts and YouTube channels, the finance industry.
- 1:02
- Unidentified speaker: We've got some receipts.
- 1:04
- Unidentified speaker: I've been watching tape on the Q3 2026 earnings call.
- 1:08
- Unidentified speaker: SK Heinex's CFO called it a virtuous cycle, where AI, like OpenAI or Anthropics Claude, starts to get recycled back into end markets, including the semiconductor industry itself, to design the latest and greatest, most powerful semiconductor products.
- 1:29
- Unidentified speaker: So what you're saying is a circular relationship is not always a bad thing.
- 1:33
- Unidentified speaker: No, all human activity is ultimately, at the end of the day, circular in nature.
- 1:39
- Unidentified speaker: And we're going to come back to another part of our semiconductor industry flow, the supply chain chart later on.
- 1:48
- Unidentified speaker: If you missed it, we have the notes on the Anthropic IPO that will also be on the midweek CSI Live.
- 1:56
- Unidentified speaker: So that'll deal with the end market, but we're going to later in this video back up to the wafer fab and packaging and assembly equipment segment as well.
- 2:05
- Unidentified speaker: It's an important spot.
- 2:07
- Unidentified speaker: All the memory companies have sold off pretty hard since the end of June, and so have their key suppliers, the companies that provide the equipment, so that these guys can go make wafers, they can make pizzas, they can slice them up into chips, package them up into a box for delivery.
- 2:26
- Unidentified speaker: Let's use Casey's pizza kitchen analogy here.
- 2:30
- Unidentified speaker: So we'll come back to those equipment makers because, as we've said for years, that is the actual enduring bottleneck.
- 2:38
- Unidentified speaker: Bottlenecks crop up all the time in manufacturing parts of the economy.
- 2:43
- Unidentified speaker: Anything that's really hardware-based, that's an ingredient to an end market product or service, most of those get solved.
- 2:52
- Unidentified speaker: That's what engineers do.
- 2:53
- Unidentified speaker: They figure out ways to get that product out when there's demand for it.
- 2:58
- Unidentified speaker: But there are actual spots in that supply chain flow where there are harder to solve for bottlenecks.
- 3:07
- Unidentified speaker: And as of right now, those still have not been solved, and we don't foresee those getting solved anytime soon.
- 3:13
- Unidentified speaker: Okay, so you'll mark that for later.
- 3:15
- Unidentified speaker: Let's just walk through some of the revenue numbers for SK Heinex.
- 3:20
- Unidentified speaker: They had a record quarter over 100 trillion won in this most recent quarter.
- 3:26
- Unidentified speaker: That's up 32% quarter over quarter, 328% year over year, operating profits up a ton as well.
- 3:34
- Unidentified speaker: 83 trillion won, up 38% quarter over quarter, and 632% year over year.
- 3:42
- Unidentified speaker: And notice on the very right portion of the screen that net income down 58% quarter over quarter.
- 3:49
- Unidentified speaker: Nick, any ideas on why that is?
- 3:51
- Unidentified speaker: Yep.
- 3:52
- Unidentified speaker: The reason for that, operating profit up quite a bit.
- 3:56
- Unidentified speaker: Net income, though, takes in non-operating income and expense.
- 4:01
- Unidentified speaker: And because SK Heinex has a big stake in Keoksia, also the manufacturing partner for Sandisk, there are some one-time equity gains and also then some one-time equity declines.
- 4:15
- Unidentified speaker: Keoksia stock has been pretty volatile from one quarter to the next.
- 4:19
- Unidentified speaker: So for a better read on SK Heinex's actual operations, you just want to back up to the middle part of the chart and just look at operating profit and not the actual net income because of those equity gains and losses tied to Keoksi stock volatility.
- 4:37
- Unidentified speaker: And if you want to see that in US dollars, here's the revenue nearly $80 billion for the quarter, operating profit of $62 billion, and that net income nearly $30 billion.
- 4:48
- Unidentified speaker: Okay, Nick, so why don't you walk us through what the revenue breakdown is for SK Heinrich?
- 4:54
- Unidentified speaker: DRAM, NANDs, where does HBM fit in and how did that all shake out?
- 4:59
- Unidentified speaker: Yeah, DRAM is the biggest part of the revenue, 75% of sales in Q3, and that is also where HBM will fall in.
- 5:09
- Unidentified speaker: None of the memory oligopoly, breakout specific HBM revenue for AI data center servers.
- 5:16
- Unidentified speaker: But you can very much think of it in very similar terms because you need DRAM wafers to make HBM.
- 5:32
- Unidentified speaker: NAND Flash at 24%, and then a very, very tiny little sliver of other revenue.
- 5:39
- Unidentified speaker: But specifically on NAND flash, a lot going there as well as the layers of NAND Flash continue to stack up.
- 5:46
- Unidentified speaker: And an important part of that work also happening with Sandisc on developing high bandwidth flash, which we've covered a number of times over the last two years.
- 5:55
- Unidentified speaker: You can check our older videos on that.
- 5:57
- Unidentified speaker: But a lot of changes and a lot of experimentation happening in memory architecture by companies that design and build these servers, and the companies that design and build data centers, the hyperscalers especially, to figure out best ways to optimize memory and long-term storage for AI applications and to ultimately improve the operations, the profitability of those very expensive assets, because ideally, they don't want to be paying this large of a percentage of their cost to build these data centers to memory and storage proprietors like SK Heinex.
- 6:35
- Unidentified speaker: SK Heinrichs also reported that HBM 4 is now in full swing.
- 6:40
- Unidentified speaker: They're sold out through 2029.
- 6:43
- Unidentified speaker: They said that the DRAM and HBM volumes through 2029 are basically completely sold out because of a combination of long-term agreements and overall customer demand.
- 6:56
- Unidentified speaker: One other key point that management said is over 50% of those long-term agreements cover the capacity through 2027 and 2028.
- 7:05
- Unidentified speaker: So we've seen this becoming much more popular with these companies that traditionally are very cyclical making these long-term agreements so the rug doesn't completely get pulled out from under them.
- 7:16
- Unidentified speaker: And we still expect that there will be a cycle at some point in the future.
- 7:21
- Unidentified speaker: It does keep getting pushed out towards the end of this decade, maybe even through the end of the 2020s and through 2030.
- 7:29
- Unidentified speaker: We'll just have to wait and see.
- 7:30
- Unidentified speaker: So the way we're still thinking of this, though, is this is above average sales cycle in length.
- 7:38
- Unidentified speaker: And of course, the magnitude of the growth in the cycle is unlike anything that's been seen since the dot-com era during the early days of the internet and infrastructure for the internet getting built out.
- 7:50
- Unidentified speaker: So pretty incredible time for companies like SK Heinex to rake in a huge windfall.
- 7:57
- Unidentified speaker: And it gives them potential, like we talked about with Micron just a few weeks ago, to evolve the business somehow and prepare long term for whatever may come next.
- 8:09
- Unidentified speaker: But as far as things appear right now, backed up by the financials in Q3 2026, SK Heinex is in good shape.
- 8:18
- Unidentified speaker: You can see this in the DRAM bit shipment growth, guided to be up at mid 20% range, quarter over quarter, driven by, as Casey just said, HBM4 in full swing, also HBM3E.
- 8:33
- Unidentified speaker: And then also just good old-fashioned DRAM.
- 8:37
- Unidentified speaker: SK Heinex, Samsung, Micron all working on cramming more bits into each DRAM chip on the NAND front.
- 8:46
- Unidentified speaker: That's something that we'll cover separately when we get an update from Sandisk and Keoxia.
- 8:52
- Unidentified speaker: Also, bit growth expected quarter over quarter, mid-single digit percentage increase quarter over quarter.
- 8:59
- Unidentified speaker: Also pretty great.
- 9:00
- Unidentified speaker: Not as big as DRAM, but really, really solid.
- 9:04
- Unidentified speaker: Let's talk a little bit more about those bit shipments.
- 9:07
- Unidentified speaker: We also said that they are in a sold-out capacity state over the next few years.
- 9:12
- Unidentified speaker: A number of analysts questioned this, asking what happens if they build too much capacity.
- 9:17
- Unidentified speaker: So let's break that down.
- 9:19
- Unidentified speaker: Let's paint it with a little color.
- 9:21
- Unidentified speaker: Should we do the difference between a Korean analyst and an American analyst?
- 9:25
- Unidentified speaker: We had actually prepared a brief segment here to compare the difference between a very polite and very orderly Korean analyst, limiting their comments to just one question and being very gracious to an American analyst.
- 9:41
- Unidentified speaker: Totally different.
- 9:43
- Unidentified speaker: Yeah, I have one question broken down into 10 points.
- 9:47
- Unidentified speaker: And I'd like you to help me connect the dots on my back of the envelope math here.
- 9:53
- Unidentified speaker: I'll stop there.
- 9:54
- Unidentified speaker: Okay.
- 9:55
- Unidentified speaker: We're going to get in trouble.
- 9:57
- Unidentified speaker: Okay.
- 9:57
- Unidentified speaker: So to actually tee up what you just said, Casey, there is more capacity coming, but the memory and storage oligopoly would like to meter that so that they don't overflood the market with too much capacity, too much supply than what is being demanded.
- 10:16
- Unidentified speaker: It's a good situation for them.
- 10:17
- Unidentified speaker: Obviously, the hyperscalers would prefer the opposite situation so they have cheap memory products again.
- 10:24
- Unidentified speaker: But there's another way that that demand can be addressed by SK Heinex and its peers, by utilizing the clean room space, the fabrication facility space they already have more effectively.
- 10:39
- Unidentified speaker: One reply that we got to one of the very polite Korean analysts on the call, very different from the Micron and other earnings calls, from SK Heinex's head of DRAM marketing.
- 10:53
- Unidentified speaker: Here is the reply from this head of DRAM marketing regarding the 4F squared next generation DRAM technology, which we're going to talk about here in just a moment.
- 11:03
- Unidentified speaker: There's a smaller cell area in this next generation model, the 4F squared versus the 6F squared.
- 11:10
- Unidentified speaker: And so what the company is doing is utilizing the 6F squared structure to the extent possible through process maturity and stable tech platform, meaning they are utilizing their existing equipment, better utilizing it and making it applicable for this new next-gen DRAM platform.
- 11:30
- Unidentified speaker: And the way they're doing that is through their EUV-based scaling and process.
- 11:36
- Unidentified speaker: In addition to continuing to improve upon 6F squared while also developing future 4F squared product, another technology roadmap, 3D DRAM, is also being worked on simultaneously.
- 11:51
- Unidentified speaker: Now, Micron has said they're going to just continue sticking with 6F squared technology and improving upon that for the time being, and then we'll eventually transition to 3D DRAM.
- 12:03
- Unidentified speaker: It's SK Heinex and Samsung kind of taking both approaches at the same time.
- 12:09
- Unidentified speaker: And the idea there being that customers can choose which technology works best for whichever application they may be working on and have need for.
- 12:19
- Unidentified speaker: But let's discuss this.
- 12:20
- Unidentified speaker: I mean, that's just like a bunch of technical nonsense for many of you viewers.
- 12:26
- Unidentified speaker: And even for some of you that are technical viewers, it's okay.
- 12:29
- Unidentified speaker: What does it mean?
- 12:29
- Unidentified speaker: What does it mean?
- 12:30
- Unidentified speaker: Let's break it down into just simple everyday language.
- 12:33
- Unidentified speaker: So let's start with this transition from 6F squared to eventually 4F squared.
- 12:40
- Unidentified speaker: What is that like?
- 12:42
- Unidentified speaker: We came up with a little explanation here for you.
- 12:45
- Unidentified speaker: It sure looks like a schematic.
- 12:47
- Unidentified speaker: And you know how I feel about those.
- 12:49
- Unidentified speaker: On the left side, we have the analogy of sub-dividing the house, increasing the bit density on the real estate that is already there.
- 12:57
- Unidentified speaker: So you can think about this if you have a house.
- 13:00
- Unidentified speaker: Let's say you want to make two dwellings out of your one house.
- 13:03
- Unidentified speaker: You want to make two apartments, or in North America, we call that a duplex.
- 13:07
- Unidentified speaker: I know that a duplex has a different connotation in Europe.
- 13:11
- Unidentified speaker: It's also a funny movie.
- 13:12
- Unidentified speaker: Also a hilarious movie.
- 13:14
- Unidentified speaker: If you haven't seen it, check it out.
- 13:15
- Unidentified speaker: It's old now.
- 13:16
- Unidentified speaker: It's very dark, but it's funny.
- 13:19
- Unidentified speaker: But hilarious.
- 13:20
- Unidentified speaker: Okay.
- 13:21
- Unidentified speaker: Being able to rent more space in your house.
- 13:24
- Unidentified speaker: So that would be on the left, that would be subdividing the house or the apartments into even smaller apartments so you can squeeze more rent out of your real estate.
- 13:35
- Unidentified speaker: Think New York City.
- 13:36
- Unidentified speaker: Yeah.
- 13:36
- Unidentified speaker: Or Seoul, South Korea, since we're talking about SK Heinrich.
- 13:40
- Unidentified speaker: That's the 6F to 4F squared transition.
- 13:44
- Unidentified speaker: The transition to 3D DRAM that might take place by the end of this decade, probably more like 2030 and beyond a type of development, is actually going to take some cues from 3D NAND, which we mentioned earlier, that continuous adding of layers in a NAND flash chip to cram in more storage capacity per chip.
- 14:08
- Unidentified speaker: 3D DRAM is doing the same thing.
- 14:11
- Unidentified speaker: This is not the same thing as HBM, where DRAM chips are actually being stacked on top of each other.
- 14:18
- Unidentified speaker: This is within one single chip itself, adding more layers, so not just one story, but a multi-story dwelling.
- 14:27
- Unidentified speaker: If you already have an apartment complex, maybe you have four apartments that all sit on the ground level, and now you start building more levels on top of it.
- 14:38
- Unidentified speaker: You gotta have rights to the airspace, though.
- 14:41
- Unidentified speaker: That's an Easter egg there.
- 14:42
- Unidentified speaker: That is a deep reference.
- 14:44
- Unidentified speaker: Either way, how this fits back into the quote from Mr.
- 14:49
- Unidentified speaker: Park, S.K.
- 14:50
- Unidentified speaker: Heinnix's head of DRAM marketing, is existing clean room space, existing fab facilities are being slowly upgraded so that these better, new and improved technologies can be run on the same existing manufacturing node.
- 15:06
- Unidentified speaker: And this is important because there are new fabs coming.
- 15:11
- Unidentified speaker: It's just not going to happen that fast.
- 15:14
- Unidentified speaker: It's not easy to build a multi-billion dollar facility.
- 15:17
- Unidentified speaker: It does take a bunch of time.
- 15:19
- Unidentified speaker: Here's a look at some of the memory fabs coming online in the next few years.
- 15:25
- Unidentified speaker: These are primarily from Micron and SK Heinex.
- 15:28
- Unidentified speaker: You can see all the way out to 2033 is when the last fab for SK Heinex is slated to get finished.
- 15:34
- Unidentified speaker: We're a ways out.
- 15:35
- Unidentified speaker: And we do have a couple of these that will begin next year in early 2027 for SK Heinex and Micron, the blue line underneath that.
- 15:47
- Unidentified speaker: But some of the more dramatic increases to capacity won't even start coming online, which isn't even to say they'll be at full capacity right away.
- 15:56
- Unidentified speaker: They'll start coming online in 2028.
- 16:00
- Unidentified speaker: So what does this mean?
- 16:01
- Unidentified speaker: In the meantime, those existing manufacturing facilities can be retrofitted with new equipment and equipment upgrades to help meet some of the bit demand growth coming from AI data centers.
- 16:14
- Unidentified speaker: And who should you automatically think of when you think of not just new clean room space that needs to be filled up with equipment, but also equipment upgrades?
- 16:23
- Unidentified speaker: Ta-da!
- 16:25
- Unidentified speaker: The Fab5.
- 16:26
- Unidentified speaker: The good old Fab5.
- 16:27
- Unidentified speaker: Especially company like Lamb Research, which has historically been focused in the memory market.
- 16:33
- Unidentified speaker: And all of them work together.
- 16:35
- Unidentified speaker: Also did the video on ASML that talked about a lot of upgrades to existing installed machinery, led to a better than expected Q3 2026.
- 16:45
- Unidentified speaker: Applied materials plays here very heavily as well.
- 16:49
- Unidentified speaker: Let's do some OG chip stock investor tech breakdown here.
- 16:53
- Unidentified speaker: A walk down memory lane.
- 16:55
- Unidentified speaker: These are provided by Lamb Research.
- 16:58
- Unidentified speaker: Let's look at how that house or existing apartment can be subdivided first.
- 17:03
- Unidentified speaker: Lamb Research said they did all this work with their simulator 3D software.
- 17:08
- Unidentified speaker: This is a side view of a one-layer DRAM chip, good old-fashioned DRAM chip.
- 17:15
- Unidentified speaker: DRAM requires power to hold the data that's in memory.
- 17:20
- Unidentified speaker: And it holds that charge in with a capacitor.
- 17:23
- Unidentified speaker: One capacitor and one transistor makes up a bit for memory.
- 17:28
- Unidentified speaker: And separating those pairs of capacitors and transistors, you have word lines and bit lines, BL and WL, they're in the middle.
- 17:36
- Unidentified speaker: Another look at that via schematic on the right.
- 17:39
- Unidentified speaker: Finally, an acronym that makes sense.
- 17:43
- Unidentified speaker: And so the challenge here is if you want to make 3D DRAM, not just improve the density of the capacitor and transistor, you also have to improve the word line and bitline dimensions as well, and make sure that the charge being held doesn't escape somewhere and lead to lower performance parasitic capacitance.
- 18:06
- Unidentified speaker: Article link at the bottom of the slide.
- 18:08
- Unidentified speaker: How do you take this from a two-dimensional structure to a three-dimensional one?
- 18:14
- Unidentified speaker: This is where the analogy of subdividing the house or building on top of your existing real estate to build a multi-story building starts to break down a little bit because what's happening here is the same schematic on the left of the capacitor and transistor and the word lines and bit lines getting tilted on the side.
- 18:36
- Unidentified speaker: And then they can stack those structures vertically on top of each other.
- 18:40
- Unidentified speaker: But that still doesn't do the trick because how do you interconnect all of those different bits in storage?
- 18:47
- Unidentified speaker: Lamb Research took it one step further.
- 18:50
- Unidentified speaker: And okay, actually, now the analogy works again.
- 18:54
- Unidentified speaker: There's the bit line there in the middle, the big fat vertical line BL.
- 18:59
- Unidentified speaker: There's the bit line.
- 19:00
- Unidentified speaker: Think of that like a central elevator, and then each stop on the floor to the left and the right, there's your word lines, there's a little hallway, and then the apartment.
- 19:12
- Unidentified speaker: So they've kind of tipped the traditional structure on its side, stuck a big elevator down the middle of it.
- 19:18
- Unidentified speaker: And this is ongoing work designing it as one thing, actually being able to manufacture it as something else entirely.
- 19:25
- Unidentified speaker: But this is why we always come back ultimately to these, not just some temporary bottleneck in the manufacturing supply chain, but why we always ultimately, as long-term investors, end up coming back to the more durable bottlenecks, ones that are not so easy to solve.
- 19:43
- Unidentified speaker: It's because Lamb Research doesn't just provide equipment for manufacturers, they actually are integral to designing and coming up with the process itself.
- 19:54
- Unidentified speaker: And then they monetize it, of course, via the sale of the machine or upgrades and other services.
- 20:01
- Unidentified speaker: But you get the idea.
- 20:03
- Unidentified speaker: This is not a bottleneck that is easily solved.
- 20:05
- Unidentified speaker: You don't just decide, LAM research, we've decided to go get equipment somewhere else, or design our own equipment, or hey, just expand your equipment capacity and problem solved.
- 20:18
- Unidentified speaker: Those equipment makers develop so much of what the industry does, so much of the design work and manufacturing work and the money flows through them that this is just a pretty durable spot in the supply chain industry flow chart.
More from Chip Stock Investor Podcast
Nvidia After Jensen: What Investors Can Do Next
Episode discussion: Speakers advise holding NVDA amid leadership uncertainty and warn against media speculation
Speakers’ claims: The hosts describe a volatile week for Nvidia investors, criticizing media outlets for presenting speculation as fact regarding Jensen Huang's potential exit. They emphasize that while charismatic leadership may be lost, operational continuity remains with interim CEO Colette Kress and long-tenured lieutenants. The speakers are holding their NVDA position without adding to it, adhering to a 'panic early' strategy rather than selling after a drop. They advise against buying the dip due to insufficient data and are instead diversifying capital away from semiconductors. Finally, they warn that many chip companies lack the staying power to survive the normalization of AI capex, urging investors to only buy dips on high-quality businesses.
Rocket Lab Is BOOMING -- So Why Aren’t We Buying RKLB?
Episode discussion: Rocket Lab Q3 2026 Earnings: Record Revenue, Neutron Delay, and Iridium Acquisition Analysis
Speakers’ claims: Speakers analyze Rocket Lab's Q3 2026 earnings, highlighting record revenue of $268M and a $3.6B backlog driven by Space Systems. The episode details the company's vertical integration thesis, including the pending Iridium acquisition which adds $900M in revenue but also $1.93B in debt. Key operational updates include the Neutron launch vehicle delay to 2027 and a 25% dilution from stock issuance to fund the deal. Speakers project full-year 2026 adjusted gross margins of ~41% and discuss the path to profitability over a 10-year horizon.