In short: Steve Weiss's highest-conviction name on the show and, he says, his largest position — initiated buy at Stifel today with a $515 target, and one of the better performers in a rough three months for the space. The thesis is the toll booth: "Jensen Huang often says if I can only get more capacity out of Taiwan Semi — and they make the chips for everybody. If you don't want a fab — and who does — they're the ones you go to." On the added capacity, he inverts his usual caution: "usually I'm a little cautious on that, but they'll fill it with demand… and it's going to take years to add it." The conclusion is the sharpest line of the hour: "to me, this is the must own in the sector. I think it's a better must own, frankly, than NVIDIA because Nvidia's getting their chips made there. As I said, so is Google, so are all the others." He has trimmed only on size — "the position got so big through performance" — not on view. Lebenthal refuses the either/or, and Wapner intervenes to make sure viewers don't hear a swap recommendation.
Taiwan Semi does not design chips; it manufactures them for everyone who does — NVIDIA, Google, Apple, AMD, Broadcom. A "fab" is a chip factory, and building one costs tens of billions of dollars and takes years, which is why almost no designer wants to own one. That is the whole business.
Weiss's claim is that this makes it structurally better than owning any single designer: "a better must own, frankly, than NVIDIA, because Nvidia's getting their chips made there. So is Google, so are all the others." Whoever wins the AI chip race, the wafers come from the same place — so you are buying the outcome of the competition rather than a bet on one competitor.
He also flips his usual caution about capacity expansion. Normally a manufacturer adding capacity is a warning sign, because new supply arrives just as demand fades. Here he argues the reverse: demand is already booked, and the physical build takes years, so the new capacity will be full before it exists. Worth noting his one trim was for position size after a run, not a change of view — a distinction that matters when reading anyone's sells.
In short: Cited as the Taiwan-war tell: "Warren Buffett got out of Taiwan Semiconductor… one of the reasons he got out of it was because he was worried about war with China and Taiwan. He was early, but when that conflict does start, it'll be a problem."
TSMC makes most of the world's advanced chips, and nearly all of that production sits on the island of Taiwan. Dowd isn't rating the business — he's using Warren Buffett's exit as a worked example of pricing geopolitical risk.
Buffett bought TSMC and then sold out unusually quickly, and one stated reason was worry about a war between China and Taiwan. Dowd's point: Buffett was early, and he missed a big AI-driven run by selling — but a company whose factories all sit in the likely conflict zone carries a risk that no earnings forecast can price. And he expects China to need an external scapegoat as its own economy deteriorates, which makes Taiwan chatter more likely, not less.
24:24He got out a little early cuz the stock went up a lot cuz of the AI bubble, but one of the reasons he got out of it was because he was worried about war with China and Taiwan. He was early, but when that conflict does start, it'll be a problem. — Super interesting point.
In short: The single best expression of the quarter's thesis — strong on both the holdings and the buys side. "Amazon and Taiwan Semiconductor are now tied as the most widely held stocks, appearing among the top five holdings of 9 of the 20 funds", and TSM was also a top-five buy at four funds, level with Cerebras. Named first in the AI-infrastructure cluster (TSM, NVDA, ASML, AMAT, MU) that with the mega-cap platforms makes up "nearly half of the top holdings listed." The lead sentence of the whole piece: "new money increasingly moved toward the rest of the AI supply chain."
TSMC is the foundry that actually manufactures the chips other companies design — NVIDIA's accelerators, Apple's processors, AMD's and Broadcom's silicon all come off its lines. It sells to every side of the AI race at once, which makes it the closest thing to a toll booth on AI compute.
In Q2 it became, jointly with Amazon, the most widely held stock across the 20 funds — a top-five position at nine of them — and was also a top-five buy at four. Being simultaneously the most-held and among the most-bought is rare; it means the funds that already own it are still adding.
This is the cleanest expression of the article's headline. If you believe AI demand is real but can't be certain which chip designer wins, you buy the factory they all have to use.
In short: Cited as a margin risk running into Nvidia rather than a position: "things that could hurt gross margins for Nvidia could be cost pressures, TSMC raising chips-on-wafer substrate packaging costs."
30:38So, things that could hurt gross margins for Nvidia could be like cost pressures, TSMC raising chips on wafer substrate packaging costs, high bandwidth memory costs, which is what hurt Apple. People are going to look at this earnings and look at hyperscaler CapEx read-throughs. A significant portion of Nvidia's revenues, right, comes from big cloud providers like Microsoft, Alphabet, Meta, Amazon.
In short: Cited as the bull case he rejects: TSMC saying demand is "insatiable" is exactly the sort of datapoint retail is chasing — "the answer is because that's known. The market is a discounting mechanism." Semis and memory remain the most crowded trade in the world.
TSMC makes most of the world's advanced chips, and it told the market demand is "insatiable." Hayes' point is not that this is untrue — it's that everyone already knows it. Markets price in what's known; the stock only moves on what's new. So when people ask how he can sell a rally backed by great chip earnings, his answer is "because that's known."
He adds a warning from history: earnings and revenue-growth expectations across the sector are the highest since late 2021, and 2022 is what happened immediately after the last time expectations peaked. Semiconductors and memory remain the most crowded trade in the world — which is why he expects them to weaken again over coming months rather than break out.
12:19seller of this move when you just saw CoreWeave's earnings and you saw Nebius's earnings and you saw Taiwan semiconductor saying that demand is insatiable and Micron and all these guys and the answer is because that's known. So what you had, you have not seen revenue growth were earnings expectations at levels this high, 15% revenue growth since Q4 of 2021.
In short: Firestone owns it and flags monthly sales up 45%: "it's incredible… there's demand, we hear about it all the time." Her demand proof is everyday AI use ("whatever, Claude, ChatGPT, every single day, building models, making charts… it's my best friend") — "you have to drive all this with chips, and the chips come from" Taiwan Semi.
TSMC actually manufactures the chips that Nvidia, Apple and everyone else design. Karen Firestone owns it and points to sales up 45%. Her demand check is refreshingly ordinary: she uses Claude and ChatGPT every day to build models and make charts — "it's my best friend" — and every one of those queries runs on hardware. "You have to drive all this with chips, and the chips come from" TSMC.
In short: The capex read-through: 2026 capex guidance raised from $52-56B to $60-64B and full-year revenue growth lifted from "more than 30%" to above 40%. But even TSMC alone can't carry Google's 2028 TPU target.
Full passage: premium transcript (PDF).
In short: Luria's first link in the value chain: "the companies that make the stuff that makes chips, primarily ASML and TSMC." Ives adds the field note from a Taiwan fab where "they're working 18 hours a day," and demand:supply for chips running "15 to 1" across his recent Asia trip.
11:39And supposedly this model is just as good as my model. So what am I going to do? You referred to AI as one business. It's not. Okay. We're talking about a whole value chain that's being created. There's the companies that make the stuff that makes chips, primarily ASML and TSMC, but a whole other slew of companies. There's the companies that make the chips, Nvidia, AMD, Micron, etc.
In short: A stellar double beat (EPS 437 vs 337, GM 67.7%, ~66% of rev now HPC/AI) — but the ADRs fell and the stock ended down high single digits for the week because capex guidance was raised 60→64B ($100B in Arizona). "You want it to give money back as it slows, but that's not happening." Taiwan's Taiex slid into a technical correction.
TSMC, which makes the world's most advanced chips, reported a great quarter (earnings and margins both beat, ~66% of sales now AI-related). But the stock fell anyway because it raised how much it plans to spend building factories (from $60B to $64B). The market's frustration: when growth slows you want a company to hand cash back, not keep plowing it into capacity. A good business, but not a fresh buy here.
Full passage: premium transcript (PDF).
In short: Santoli's market color: another "phenomenal AI hardware type result" (with Micron and ASML) that drew a "pretty poor market reaction" — the series of good-numbers-still-sell prints defining the momentum flush. No individual call.
In short: Cited by Lebenthal as the week's first near-term read on the semi space — Taiwan Semi reports at the end of the week. Context for the memory/semi debate; no individual call.
In short: Reports this week alongside ASML and "expected to do phenomenally well." Carlson is "still very bullish fundamentally on ASML and TSM," but is "not making any big bets on them this quarter" — trimming/shifting toward quality-at-low-valuations rather than piling into semis as he expects the trade to slow.
TSM reports alongside ASML and Carlson expects a great quarter — he's "still very bullish fundamentally." But like ASML, he's "not making any big bets on it this quarter." After a huge run in the whole semiconductor group, he'd rather add to cheaper, out-of-favor quality companies than pile more money into semis right as he thinks the trade is about to slow. So it's a hold-and-admire, not a fresh buy.
21:00So, while I'm still very bullish fundamentally on ASML and TSM, I'm not making any big bets on them this quarter. They could go up 10 20%, we could see another bit of momentum in them. I'm sure the reports are going to be great, but I'm actually in the position right now of trimming and shifting the portfolio towards the quality growth companies at low valuations than I am piling more into semiconductors.
In short: Named as a Veolia water-services customer — Veolia is working with TSMC at its mega chip complex outside Phoenix. Cited to illustrate the data-center / advanced-manufacturing water opportunity, not as an investment view.
In short: Giroux: named a diversity-of-compute winner (it fabs everyone's silicon — GPUs and the ASICs displacing them alike). Rossbach flags TSMC-centered Asian supply-chain pressure as the main risk to Arista's ramp — a read on how tight leading-edge capacity is.
In short: Weiss's resolution to the messy semi/memory debate: "all roads lead to Taiwan Semi" — whether it's Meta, Apple or NVIDIA making the chip, they all need TSMC capacity. The cheapest and most predictable semi because it doesn't price off volatile spot markets: "you want their capacity, you pay for it, and that's it."
Taiwan Semiconductor is the factory that actually manufactures the advanced chips designed by Nvidia, Apple, Meta and others — nobody can make leading-edge chips without it. Steve Weiss's point is that however messy the memory debate gets, "all roads lead to Taiwan Semi": everyone needs its capacity, so it captures value no matter which chip designer wins.
He also likes that it's more predictable than most chip stocks: it doesn't price its output off volatile spot-market prices — you reserve its capacity and pay a set price — which makes its earnings steadier and, in his view, makes it the cheapest and most reliable semiconductor to own.
In short: Allocation context — while the industry fights over scarce TSMC capacity and HBM, Cerebras sidesteps the queue: no high-bandwidth memory, no CoWoS advanced packaging, no bleeding-edge 3nm, leaning on plentiful SRAM and mature 5nm. Referenced/neutral.
In short: ASML's primary customer — "where all the action is," up ~50% YTD / ~122% on the year. Part of the parabolic semi vertical he's rotating away from (not selling, but not buying).
8:34We can look at the other obvious example, ASML's primary customer, which is TSM. TSM is in the green today. It continues to go up. Like ASML, it's up 50% this year. Over the past year, it's up 122%. TSM is where all the action is. People love this bottleneck stock today.
In short: "The most conservative people in this ecosystem" — TSMC (with the Koreans and ASML) constrains capacity, keeps Jensen from flooding the market with Vera Rubin chips, and has prolonged the AI cycle by a couple of years.
TSMC manufactures essentially all the world's cutting-edge AI chips. In a market full of cheerleaders, Singh calls the Taiwanese (with Korea's chipmakers and Holland's ASML) "the most conservative people in the ecosystem": they refuse to recklessly expand capacity, which stops Nvidia from flooding the market with chips.
That discipline is, in his view, what has kept the AI boom from overheating and bursting — it has stretched the cycle out by a couple of years. It's the same role OPEC plays in oil: the supplier who controls the spigot ends up controlling the cycle.
38:55whether it's CEOs of Micron I think the most conservative people in this ecosystem are effectively the Taiwanese like the TSMC's of the world. and the Korean chipmakers and ASML the Dutch lithographer the lithography company manufacturer and they're the ones constraining capacity in the market.
In short: Referenced with ASML/Nvidia as a "bottleneck" name the market is crowding into.
15:20Well, I do and I still rate it as a buy. One thing you'll notice with all of these companies that have moved down over the past year is they're not in the new exciting thing. The new shiny exciting thing is artificial intelligence and especially the bottleneck stocks. The ones like ASML, Nvidia, TSM, any company that has a bottleneck. You have memory stocks.
In short: "Really levered to it" — a core AI winner in his new Mega-Cap Tech Fund.
TSMC is the world's dominant chip manufacturer — almost every advanced AI chip is physically made in its factories. Sacerdote calls it "really levered" to AI: as AI chip demand explodes, the demand flows straight through to TSMC.
It's a core holding in his new Mega-Cap Tech Fund, which picks the best of the world's largest tech companies — names with wide moats that he thinks are structurally underweighted by investors.
1:13:31we created the Whale Rock Mega Cap Tech Fund which is the top 30 the universe is the top 30 market caps globally and then we pick you know the 12 or 13 that are the best and I think there's tremendous alpha in the largest cap... these companies by definition have wonderful modes... I mean, Nvidia sure is, and TSM is really levered to it, and Heinix is extremely levered to it, and ASML is levered to it.
In short: Cited as the leading-edge logic supply Nvidia has locked down — leading-edge logic and CoWoS packaging are "always tight," but Jensen "saw this coming" and secured supply across the value chain.
13:52— Mean supply from like Taiwan semiconductor. Well, Ty and also the memory guys and also the the packaging and now we know he's he's getting into the optical and and scing like lasers and all kinds of other stuff, right? Um but he's been very good at at at doing that. So they have supply. Um but this divergence has been very interesting.
In short: "One of the best trades in AI" — own the bottlenecked input, not the chip designers. Sole manufacturer of all seven Vera Rubin chips (3nm); "everything starts with TSMC." Morgan Stanley Taiwan checks upbeat (Rubin demand, Blackwell re-accelerating); supply relief unlikely before 2028.
TSMC is the world's most advanced contract chip factory — it physically manufactures the chips that companies like Nvidia design. It is the only place that can make all seven chips in Nvidia's next-generation "Vera Rubin" AI system.
The idea is to own the bottleneck rather than the chip designers: "everything starts with TSMC," and there's no rival that can relieve the supply crunch before about 2028. That scarcity is why it's called "one of the best trades in AI."
Full passage: premium transcript (PDF).
In short: Cited as a funding-stress example, not a stance: prime brokers were charging 3%+ to fund TSMC (and Hynix/Samsung/Kioxia) longs — abnormal, since longs normally earn a credit from PBs who rehypothecate the stock.
TSMC is the world's dominant contract chipmaker. It shows up here only as evidence, not as a buy or sell call. Normally, when you hold a stock through a prime broker, the broker lends your shares to others and pays you a small credit for the privilege. Paulo's contacts say the opposite was happening: prime brokers were charging 3%+ just to let clients stay long TSMC (and Hynix, Samsung, Kioxia). Paying to be long instead of being paid is a clear sign that demand for leverage in these crowded Asian-semiconductor names had overwhelmed the supply of financing — a funding-stress tell, not a view on the company.
In short: "A very durable seller," not highly cyclical — leading-edge foundry capacity, "one of the most important companies on planet earth." A valid phase-1 pick alongside ASML.
8:18ASML still have a little bit of cyclicality. Then, of course, we have TSMC. TSMC is a very durable seller. It's not some highly cyclical stock. We have Nvidia which I believe because of CUDA, because of the infrastructure, because of the ecosystem, it has some cyclicality, but I truly believe that Nvidia is becoming less cyclical over time.
In short: The "flinty old men and women" guarding Morris Chang's legacy won't expand as fast as Jensen wants (maybe +5% vs his "double or triple") — enforcing the real-world wafer constraint that helps everyone avoid a bubble.
TSMC is the factory that physically manufactures almost all the world's advanced AI chips. Baker's whole "no bubble" thesis runs through it: the people who run TSMC are disciplined veterans who see themselves as guardians of founder Morris Chang's legacy and protectors of Taiwan, so they deliberately refuse to expand capacity as fast as customers like Nvidia demand.
That restraint is a real, physical limit on how fast the AI build-out can grow — a brake that didn't exist in past tech manias. By rationing the supply of chip-making capacity ("wafers"), TSMC keeps the boom from overheating into a classic bubble. That makes it both a beneficiary and the safety valve.
10:46And the reason it's going to persist for a long time is Taiwan Semi is run by flinty old men and women in their 70s. Not to say 70s old, it's the new 50. I'm 50, it's the new 30. — [laughter] — But they're the most important people in Taiwan. They are Taiwan. And they view themselves as the guardians of Morris Chang's legacy.
In short: A top holding, reframed as industrial/infrastructure — "a factory of one," probably the most valuable piece of infrastructure the world has (a fab is like an aluminum smelter or copper mine, GICS code aside). The cash flows here; will keep doing incredibly well.
TSMC makes the world's most advanced computer chips. He insists on seeing it not as a tech stock but as heavy industry — a "factory of one," in the same bucket as a copper mine or aluminum smelter, just with a misleading industry label.
It's one of his biggest holdings and, in his view, probably the single most valuable piece of infrastructure on the planet — if anything happened to its Taiwan plants, "we're all freezing in the dark." It's exactly where all the money is flowing, so he expects it to keep performing.
15:36build all of this out. Basically, [clears throat] study the supply chains, and you brought up semiconductors. I view Taiwan Semiconductor, just to use that as a proxy for the semiconductor industry, I view that as an industrial or infrastructure company. I mean, it's effectively a factory, and it's a factory of one, and they make some of the most valuable products in the world.
In short: The lone foundry spender for a decade that "underspent significantly" and "made a mistake… they will have to rectify." High-margin (~70%) but framed as the customer whose under-investment + new competition sets up the multi-spender equipment boom, not a single-name buy.
TSMC is the world's dominant contract chip manufacturer ("foundry") — it makes the advanced chips designed by Nvidia, Apple, and others. In Shaulov's story it's less a buy than the setup for his real trade. For a decade it was essentially the only big spender on new capacity, and he argues even it "underspent significantly" — "made a mistake" it will have to fix.
That under-investment, plus new foundry competition from Intel and Samsung, is exactly what creates his thesis: multiple chipmakers all needing to spend at once, which is great for the equipment suppliers. TSMC's ~70% margins show how much room there is to spend.
10:25But most of it is just on an incredible trajectory. If you look at AI demand, it's driving so much compute demand. And first we started with GPUs, then we went to memory, now it's CPUs, networking chips. And that is just creating massive supply constraint. And if you think about the semiconductor industry — maybe just take the last decade — they've gone through so many boom and busts that most of these companies have gotten pretty disciplined about CapEx. I.e.
In short: Named once, as ASML's customer base: "Every major chipmaker in the world, from TSMC to Intel, buys its machines from ASML." No stance.
In short: Named once, as Scottish Mortgage's second-largest position at 5.7%. No stance.
In short: #5 on the Q4 2025 table and #7 over two quarters. Aggregate-flow data point; no view.
In short: Featured as a profit-taking case study, not a fresh buy — first flagged ~$84 (Aug 2022), bought the $70 dip, trimmed 25% at a time up to ~$303 (a "four-bagger"; the DCA-out sim returned 215% on the original $12,500, ~46% IRR). Verdict today: "no longer the bargain it once was" — P/E ~20 is okay but 13× sales is "way up there" and "leaves scant room for disappointment." Dominant in precision wafer etching (essential to Nvidia's AI chips) but richly valued: exit the rest or do another partial sale; the "elderly Haymaker is retaining a modest position." A trim/hold, not an add.
This week isn't really a "buy this" pick — it's a lesson on selling, using TSMC as the example. TSMC is the Taiwanese company that physically manufactures the world's most advanced computer chips (Nvidia designs AI chips; TSMC actually etches them onto silicon). Haymaker recommended it years ago around $84, and it has since quadrupled to ~$303. Rather than sell it all at once or never sell, Haymaker's approach is to take a slice off — about 25% of the position — each time it runs up a lot. The article works through the math to show that this "trim as you go" method captured almost the same total gain as never selling, but with far less risk of giving it all back if the stock crashes.
On TSMC specifically, the message is: it's no longer cheap. Its price is now 13 times its annual sales (a very rich level; price-to-sales compares the stock price to revenue), which "leaves scant room for disappointment." So if you own it, either sell the rest or take another partial profit (Haymaker himself is keeping just a small position). The same warning applies to two other past winners that got expensive — Alphabet/Google (GOOG) and ExxonMobil (XOM) — and the bigger-picture advice is that with so many stocks looking pricey, investors should raise a meaningful cash cushion of 20–25%.
In short: UI demonstration only — no view expressed. The name that happened to be sitting in the dashboard's notifications panel during the demo: "Taiwan Semiconductor just posted this event. Okay. Let's look at what that is. Oh, it's their Q4."
1:09:17Taiwan Semiconductor just posted this event. Okay. Let's look at what that is. Oh, it's their Q4. Okay. Awesome. Interesting. Let's read about this and let's figure out what the company just reported. And so this is a home base for a very personalized investing experience, is the dashboard. In terms of what's coming —
Nothing matches this filter.
Verbatim excerpts from the public transcripts (auto-pulled at each mention's timestamp, lightly cleaned). Timestamps link into the video; "source page" opens that commentator's full analysis at this row. Click a mention's header line to expand it (one open at a time). For personal study — not investment advice.