ASML Outlook Observations, TSMC Q2 Recap, Intel & Samsung Culture Woes, More on Mainstream
Chips & Wafers Weekly Update
ASML Outlook Observations
ASML 0.00%↑ announced Q2 results earlier this week, missing Q3 revenue outlook (7.65 EURb Vs. Est 8.19 EURb) while showcasing strong bookings toalling 5.5 EURb (Vs. Est 4.44) of which 2.3 EURb are for EUV tools.
Despite the strong orders, the stock dipped as the trading day progressed.
ASML is an extraordinary company, with their tools being used for 100% of all the worlds’ advanced IC’s.
However, company excellence & market dominance don’t always mean strong stock performance.
Here’s our macro view on ASML:
ASML indeed holds a monopoly on advanced chipmaking efforts globally. However- this monopoly and dominance IS ALREADY PRICED IN. The monopoly won’t change likely in the near future, but the market already knows that. This means that the monopolistic nature of the company is somewhat irrelevant to the stock performance discussion, which is then diluted to two main points: (1) Advanced tool bookings (EUV & High-NA) for both logic and memory, and (2) inflated DUV sales for the insatiable mature-node capacity expansion efforts in China.
For the first point, advanced tool orders for logic are still somewhat soft comparative to historical, with memory accounting for lower and lower percentage of total bookings.
As for the second point, China revenue continues to stay in the 25% range QoQ after nearing nearly 50% (!) just last year - as Chinese stockpiling efforts of 2024 continue to decelerate. Another ominous sign for ASML’s long term outlook.
Lastly, interestingly, there may actually be some cases where market competition is better for stock performance over monopolies. How so? As mentioned, sustained monopolies are priced in, competitive advantages on the other hand, are not, because they can change.
So when TSM 0.00%↑ announces stellar earnings with robust outlook and continues to demonstrate market leadership ahead of INTC 0.00%↑ & Samsung, that is a new market surprise every quarter that is not guaranteed, but rather earned - very unlike ASML’s monopoly position.
To conclude, we believe ASML is one of the most extraordinary companies in the world, but are not as optimistic about how that reflects in their stock price in the near term.
TSMC Q2 Recap
What a company.
Let’s recap the quarter, outlook, and the point out key comments from our perspective:
Q2 Results:
Revenue: $30.07 B (Est. $30.00 B), +44.4% YoY
Net Income: $13.5 B (Est. $12.1 B), +60.7% YoY
EPS: $2.47 (Est. $2.12), +60.7% YoY
Gross Margin: 58.6%, +5.4 pp YoY
Wafer Shipments: 3,718K, +19% YoYQ3 Outlook:
Revenue: $31.8B–$33.0B (Est. $31.72B)
Gross Margin: 55.5%–57.5% (Est. 57.2%)
Operating Margin: 45.5%–47.5% (Est. 46.9%)
FY25 Outlook:Revenue Growth Raised: +30% YoY to ~$117B (prior mid-20%)
CapEx: $38B–$42B (Unchanged)
There is a lot to cover in this call, and maybe we’ll do a deep dive next week. But here are some key themes and comments that we found particularly interesting:
FX Gross Margin Impact:
Company commentary has said that every 1% appreciation in the TWD:USD rate translates into 40 BP of reduced GM. The TWD appreciated ~7% in Q2 Vs. Q1 implying an expected ~280 BP GM hit, yet actuals were only -20 BP (58.6% Vs. 58.8%).
This shows tremendous ability on TSMC’s side to pull multiple alternative levers (primarily wafer ASP’s?) to ensure their margin targets despite challenging Forex environments. What a tremendous capability.
Advanced Node Tool Commonality:
CC noted during the call that TSMC’s advanced nodes (7nm and below) have roughly 85-90% tool commonality - meaning that underutilized tools for 7nm can be quickly repurposed to satisfy 3nm demand or the upcoming 2nm ramp.
When thinking of investment ideas that track TSMC’s node migration, this is key to take into account. The focus areas should be the remaining 10-15% that have to be incremental WFE processes that old tools can’t execute, or alternatively, increased intensity or reduced WPH throughput that will increase the total amount of incremental tools needed in addition to existing old ones. Maybe in our deep dive next week we’ll provide some equities that play nicely into this concept.
On-Device AI Die Size:
CC made a ridiculously important point regarding on-device AI silicon: Although it’s hard to predict unit volume impacts at this moment in time, on-device AI chips are roughly 5-10% larger than standard.
What does this mean? Let’s do an abstract back-of-the-envelope calculation: Say you have smartphone SoC’s that are 90mm2, now they are 100mm2. Instead of fabbing ~600 DPW, you can now only fab ~450 DPW; a -25% reduction. So if I need to fab wafers for 1 million phones, with flat unit volumes I will now need to process ~+33% additional wafers comparative to previous generations.
This means TSMC is looking towards more capacity constraints, and WFE players that are exposed to wafer-count processing are poised to benefit.
The one thing missing for us on the call was more information on the Advanced Packaging efforts - revenue and margin profiles comparative to foundry as well as applications and timelines for new technologies such as their recently announced System on Wafer (SoW) architecture. Maybe next time.
Intel & Samsung Culture Woes
If there is one thing that has been proven true time and again in business, it is that culture is no less important than capability.
This past week we have seen concerning commentary on INTC 0.00%↑ and Samsung regarding company culture which seem to be hurting both organizations respectively.
In a recent Chosun Piece, we saw some troubling comments about Samsung, and how their toxic culture is pushing engineers to their next-door competitor SK Hynix - here are some excerpts just to get a taste:
…At Samsung, cross-departmental projects often revealed defects in process or design, but these were routinely downplayed to avoid internal blame. “Everyone knew the game,” he said. “Minimize the problem or your team gets burned.”
…Samsung-to-SK hynix moves have surged in recent years. “It’s not just about salary, It’s about rigid hierarchy, performative politics, and an HR system that discourages risk.”
…SK hynix encourages competition between teams pushing rival packaging technologies—such as MR-MUF and hybrid bonding—then selects based on performance and cost. At Samsung, he said, the dominant concern is who will be blamed if something fails.
…Experts say the issue at Samsung isn’t technology—but structure. “It’s a management-first organization,” said Kim Yong-jin, a professor at Sogang University’s School of Business. “Engineering decisions are filtered through financial metrics. If leadership doesn’t reset the system, Samsung’s semiconductor edge could continue to erode.”
Samsung is not alone. They have a friend across the ocean in the US - Intel.
We scoured through LinkedIn to search for some comments from ex-Intel employees about their experience there - no less damning. Check out this particular post after the reports regarding Intel ditching their glass core substrate efforts:
And the comments? Ruthless:
When you trade visionary engineering leadership for bean counters this is always the result!
…We (Intel) missed some BIG things like Internet, cell phone, because we did not follow Noyce’s statement to DO something; I hope Intel is wise enough now not to miss the big picture because it focuses too much on small pictures.
And check out this comment from a former director of Intel’s test division that after 27 years(!) moved to Technoprobe:
Culture matters. Employees want to work for winners and risk takers. Elon Musk is an almost tyrannical leader but he attracts talent and dedication like no other. People search for purpose, and it seems that some of the semiconductor incumbent behemoths are forgetting this.
More on Mainstream
Those who’ve been following us know that we have been arguing for a few months already that the recovery in mainstream semi has already begun, and we are at the precipice of a dual-engine surge of AI/HPC demand coupled with high-volume mainstream demand. Here’s some additional data points to support our thesis:
TSMC Q2 2025 Smartphone Revenue: Grew to $8.1 USDb, up +18% YoY.
Taiwan PCB June Sales: Total PCB suppliers for consumer electronics grew +27% YoY.
Taiwan ABF June Sales: Total ABF sales (primarily PC exposure) grew +34% YoY.
We believe it has begun, and that the next couple years are going to be very exciting for the semiconductor industry driven by this dual-growth opportunity. Exciting times!
That’s a wrap for now. See you next week!
Chips & Wafers Team
So ASML's monopoly is priced in, just like Visa an mastercard's, spgi, etc for the last 10 years.
I'll take it.
Also, how is tsm not a monopoly since a few good years?