Taiwan Semi builds chipsets for other companies like Nvidia, AMD and Qualcomm. It does so in its highly expensive, highly complex chip fabrication plants. These are called “fabs” for short and represent what I view as one of the deepest competitive moats on the planet.

Needed Definitions:

Taiwan Semi missed revenue estimates by 0.7% but beat its revenue guidance by 0.5%. Its 23.6% 2-year revenue compounded annual growth rate (CAGR) accelerated compared to 13.1% last quarter and 18.0% 2 quarters ago.

On gross margin, there were a few puts and takes to discuss. The ramping of its production in Japan and Arizona is dilutive to margins. It takes time for new facilities to reach the level of scale and maturity seen at its Taiwan facilities. Its supply chain in Taiwan is extremely advanced and efficient, which means margins there are better. Still, despite the absolute hit to margins from international expansion, it expects its margin profile to be best-in-class anywhere it operates. Meaning? The relative cost advantage it enjoys vs. the field will remain intact.

There was also an unplanned gross margin headwind during the quarter. Earthquakes in Taiwan led to significant production disruption and some wafer materials in process had to be discarded altogether. Still, through excellent execution (what else is new), the company was able to minimize the headache and quickly recover lost production levels. This allowed it to beat the midpoint of its revenue guidance. And it still beat its gross margin guidance by 80 bps despite this 60 bps headwind.

“We remain committed to a sustainable and steadily increasing cash dividend per share on both an annual and quarterly basis.”

CFO Wendell Huang

For Q2, revenue guidance beat estimates by 5.9%, GPM guidance met estimates and EBIT guidance beat estimates by 6.3%.

For the full year, it reiterated annual revenue growth guidance of roughly 25%. It also continues to expect 100% Y/Y AI accelerator growth for 2025.

Importantly, guidance includes the expected impacts of China Hopper 20 (H20; Nvidia GPU built for that market) export bans. For context, China represents about 10% of next-gen chip shipments for TSM. Reiterating annual guidance despite this is because demand everywhere else is that good. And furthermore, there was no change in customer buying patterns stemming from the threat of tariffs. The strong Q2 guidance is not a matter of pull-forwards. Some thought that would mean full-year guidance would also be raised, but I find a reiteration in this environment to be both responsible and impressive. There’s little incentive to get aggressive with forecasts as the world rapidly evolves.

Q2 gross margin guidance reflects an 80 bps Q/Q decline due to the ramp-up of its new facilities in Japan and Arizona. It reiterated expectations of this international expansion fostering a 250 bps gross margin headwind through 2025.

As a reminder, it also announced a new $100B investment in Arizona to support construction for a few more fabs (more later). Based on this, the company offered new multi-year gross margin guidance. It expects the 250 bps headwind to rise to 350 bps towards the end of the roughly 5-year planning period. That widening is also partially related to expected tariffs. Still, considering this quarter’s gross margin was nearly 59% and its long-term target is 53%, this is a move it can easily make while giving its customers more flexibility, cutting geopolitical risk and still exceeding margin goals.

TSM trades for 14x forward EPS and 27x forward FCF. EPS is expected to compound at a 2-year clip of 24%; FCF is expected to compound at a 2-year clip of 16%. The multiple is almost surely related to rising risk of Chinese invasion in Taiwan and the lower multiples that companies from that nation generally receive. Estimates likely will modestly rise following this report & guide.

Revenue Breakdown:

Advanced technology represented 73% of total revenue vs. 74% Q/Q & 69% Y/Y. This is the bucket of demand that includes GenAI, agentic AI and, more generally speaking, high-performance compute (HPC) buckets. HPC rose 7% Q/Q to reach 59% of total revenue vs. 53% Q/Q & 51% 2 quarters ago. As you can see below, 3 and 5 nanometer technologies are currently the dominant growth drivers.

By technology:

By platform:

Product Roadmap – N2 and A16 (1.6 millimeter) timing:

Leaders of rapidly growing industries must innovate with an obsessive, borderline-paranoid mindset. They must always assume the competition will catch and pass them if they don’t work as efficiently as they can. Even for TSM, with one of the more expensive and entrenched competitive moats, this is true.

For its new N2 technology, TSM expects tapeouts (customer chip designs) to exceed all technological predecessors. N2 delivers 25% better power at identical speed vs. its most advanced N3 iteration. It also boosts chip density by 15%, allowing partners to pack more into each individual chip. N2 is on pace to ramp scaled production by the end of this year. A16 will come after that. As a reminder, this features a super power rail (SPR). This is a power source that attaches to the back-end of a chip, which, per TSM, offers better density and design flexibility. It also features a 9% speed improvement or 17.5% power improvement vs. N2P. For TSM, this is a marathon and a sprint.

“We believe N2, A16 and their derivatives will further extend our technology leadership position and enable TSMC to capture the growth opportunities well into the future.”

CEO C. C. Wei

CoWoS (already defined):

TSM plans to double CoWoS capacity to support AI accelerator demand. The team fielded questions about rumored order reductions, but they’re not seeing that. No cancellations. It is confident that it needs all of this capacity to reach a better supply equilibrium by next year.

Sector Growth & Market Share within its Expanded TAM Definition:

As a reminder, TSM expanded its definition of the foundry industry to “Foundry 2.0.” This newly calculated total addressable market (TAM) includes wafer production and now includes packing, testing and mask making. It reiterated expectations for 10% industry-level growth in 2025, thanks to thriving HPC performance and “mild recoveries in other end markets.” Based on full-year guidance, TSM expects to comfortably outgrow its industry and continue taking significant market share.

More on Arizona Expansion:

Taiwan added a new $100 billion U.S. investment commitment during the quarter. This brings its total investments and plans here up to $165 billion. It will mean 30% of its N2 and more advanced production will happen in the U.S. upon completion, but this will take a few years, as the next phase of building likely won’t start until 2027. Customers want “geographic flexibility” and Taiwan Semi is giving it to them, while sidestepping potential tariffs looming for the semiconductor industry. They’re trying to speed these timelines up as much as they can, given elevated demand for U.S.-made chips. The new U.S. expansion plans add 3 planned wafer factories, 2 advanced packaging facilities and a “major R&D center.” This joins the existing roadmap of 3 advanced wafer factories in Arizona.

From the first $65B investment announced, the first of 3 plants is already in scaled N4 production as of the end of 2024. Yields are similar compared to Taiwan, but margins are still lower due to the less mature supply chain and economies of scale. The second facility will feature N3 tech and is now complete. It is in the early stages of ramping production to cater to AI demand. The third facility will deploy its newest N2 and A16 processes, with plans for construction starting by the end of this year. Considering how easily Nvidia was able to secure permits and licensing for its own planned U.S. projects, I don’t anticipate TSM having issues or delays with regulators here. If anything, they’ll be ahead of schedule — and that’s the goal. Interestingly, TSM is NOT participating in any joint ventures on these plants. They’re doing it all themselves (with the support of local and federal governments).

Japan, Europe & Taiwan:

Tariffs:

The company has seen no changes in customer demand from tariffs. This is why full-year guidance was reiterated. At the same time, it acknowledged heightened levels of uncertainty, and told investors it would have a much better sense of the impact in a few months. Perhaps in relation to all of this trade war drama, it boosted N4 pricing in the USA by 30%. It has also consistently spoken about “Made in the USA” labels fetching a sizable expected premium, so this could also be related to that.

DeepSeek:

At this point, DeepSeek driving model cost deflation is broadly seen as positive for all players supporting GenAI proliferation. Just like in cloud computing, when cost deflation directly supported access to greatly offset weakened workload pricing power, the same is expected to happen in the world of AI. That’s great news for TSM, as it means more demand for its fabrication, packaging and other services. Leadership reiterated this idea during the prepared remarks.

Really good quarter and guidance. The lack of a full-year raise doesn’t bother me, as it’s very wise to remain prudent amid the chaotically moving macro backdrop. This is the clear fab leader for AI accelerators and that will not change. The 20% CAGR at this scale is phenomenal and so is its pace of product improvement. There’s really nothing to pick at here. From a micro viewpoint, this is an elite company trading at a highly compelling price and benefitting from the AI infrastructure tailwinds raging through markets. From a macro viewpoint, the threat of a Chinese invasion continues to loom, and that overhang probably won’t go away any time soon. Furthermore, the GenAI cycle will not last forever, and timing is highly uncertain at this point. If you’re comfortable with those 2 risks, I fully understand why you are eager to own this.

Netflix needs no introduction.

Content and marketing expense timing aided the profit beats. This will mean margins during the first half of the year are higher than during the second half.

Despite the Q1 beat and Q2 revenue guidance being about 1% ahead of expectations, it reiterated annual revenue expectations. Just like for TSM, this is wise amid the global economic chaos. Netflix is about as insulated from trade wars as any other company, but nobody is 100% immune. Furthermore, we got more commentary pointing to this guide being overly prudent, if anything. The company acknowledged that FX favorability has it tracking above the midpoint of the current annual revenue guide. It just didn’t want to rely on that continuing to meet guidance. To me, all of this effectively de-risks the annual guide:

Barring an economic collapse (not at all my expectation), they should be set up well for 2025 beats and raises.

One more note on full-year guidance. Netflix reiterated expectations for 100% Y/Y ad revenue growth, modest revenue per member growth from price hikes and healthy subscriber growth.

For Q2, revenue and profit guidance were both comfortably ahead of expectations.

I expect profit estimates to rise following this report. The company trades for about 39x forward earnings before revisions and likely closer to 35x after they take place. EPS growth estimates for this year will likely rise from 24% to closer to 30%. Next year, EPS is expected to grow by 23%.

Consumer Response to the Trade War Thus Far:

“We're paying close attention to consumer sentiment and where the broader economy is moving. But based on what we are seeing right now, there's nothing really significant to note… retention is stable and strong. We haven’t seen changes in plan mix. Our most recent price changes have been in line with expectations. Engagement remains strong and healthy. So things generally look stable from that lens.”

Co-CEO Greg Peters

“We're seeing strong, stable acquisition and retention trends in the business generally.”

CFO Spencer Neumann

Leadership reminded us that entertainment is generally resilient across cycles, the Netflix subscription is an affordable luxury and best-in-class engagement inspires best-in-class loyalty. And advertising should create another resilience tailwind for any potential future economic weakness. During the last macro downturn, Netflix didn’t have the option to offer users a cheaper plan instead of canceling entirely. Now they do. That should help convince more people to stick around in any environment.

“Historically, in tougher economies, home entertainment value is really important to consumer households, and Netflix is a tremendous value in absolute terms and certainly in competitive terms.”

Co-CEO Ted Sarandos

This is all encouraging, specifically from a Netflix perspective. More generally speaking, the company is seeing no signs of softness in advertising marketplaces. It’s actually “seeing the opposite.” Despite Netflix still being very small in ads (which means market share gains can more powerfully offset any macro headwinds), I still found that encouraging for the streaming advertising landscape as a whole. And that makes sense. During times of chaos, marketers flock to their highest quality impressions. Streaming offers 100% identification rates and an easier means of tapping into programmatic advertising for more granular targeting than most other channels.

Rounding Out the Content Offering:

For streaming subscriptions, maximizing the instances of “cancellation objectors” is the best way for a company to minimize churn. If Netflix has scripted dramas, horror films, reality flicks, kid shows, stand-up comedy etc. then they can cater to a more diverse set of needs and raise the probability that a member of a household will not allow a cancellation. If I have one of those things? Subscribers will more routinely sign up to watch a title and then churn until it comes back. Simple enough.

Netflix is best-in-class when it comes to retention. And while that’s impressive, it becomes even more impressive when we consider it’s just now entering live sports. That is among the most powerful sources of customer acquisition, loyal engagement and, therefore, cancellation objections. It should merely improve its already impressive churn rates. The same can be said for gaming and the company’s more aggressive push into children’s TV via licensing Ms. Rachel. And they have not even entered the world of live news or podcasts, which represent more opportunities for more engagement and higher retention. In many industries, we can subjectively argue over which competitor is the highest quality. I think it’s obvious within streaming that Netflix is one of one, and they have incremental levers to pull to distance themselves further from the pack.

2025 should be more of the same. In Q1 it already saw 4 titles (across a wide array of genres) reach its most watched lists, with several crossing 50M total views. Love is Blind is now the most popular unscripted streaming show ever and has already spun out into local shows in 8 countries. 3 more are coming soon. That’s a great way to create more entertainment with diminished execution and consumer interest risk. This is simply leveraging a proven concept, which is what firms like Disney lean heavily on to keep producing popular titles. Netflix will inherently get more and more of these opportunities as the company gets older and keeps producing hit shows.

Squid Game and Big Mouth finales earmarked for 2025 are just 2 of several highly-compelling examples of more beloved content coming to Netflix soon. And for Squid Game specifically, the related update to its popular video game and live experience expansion in France both offer significant promise.

Competition:

Aside from beating the competition with great shows and an increasingly diverse offering, it continues to meaningfully prioritize local team and content growth across the globe. Like any other leader, it knows it must keep sprinting to stay ahead. That’s the plan… with an added emphasis this quarter on “programs for tastes, cultures and languages that appeal to local audiences.” This isn’t just about spending more money, but getting more talent in place with an expert-level understanding of what local markets want. The things that excite consumers in India are different than in Sweden… are different than in Nigeria… are different than in Brazil. You get the idea and Netflix does too. It has been investing in this international foundation to grow everywhere, knowing that localized content better serves specific cohorts and can periodically transcend borders to reach global fandom. Examples like these are bountiful.

Netflix thinks it’s the combination of great content, locally relevant programming, the quality recommendation engine and great language dubbing that position it to keep thriving everywhere. And it’s not like it has to pay tariffs to export content from one country to another.

This quarter, it offered The UK as a promising case study. Continued investments in that relatively mature market led to screen time market share rising from 8% to 9% Y/Y. It still trails BBC and ITV, but the data (from “Barb”) includes both linear and streaming traffic. Netflix has no linear business, so this isn’t apples-to-apples. If it were, it would paint Netflix in an even more positive light.

Monetization & More on Ads:

Price hikes went exactly as planned, with the churn reaction modest. It raised pricing in France today, which was already in 2025 guidance.

The company is on track to “reach sufficient scale with the member base in all ad-supported countries” this year. To make sure advertisers can access these eyeballs with great value, it’s prioritizing the usability and efficacy of its ad-buying services. Netflix Ads suite (the in-house ad-tech platform) expanded to the USA this month. This gives brands more flexibility to use their preferred partners, granularly target and access needed data. The launch has performed in-line with expectations – just like its launch in Canada did a few months ago. It will debut in its other 10 ad markets this year, and has a multi-year roadmap for improving buyer utility. There were no updates to partnerships with Alphabet or Trade Desk, which is good news for both, considering all recent headlines point to those relationships deepening.

It also launched programmatic buying processes in EMEA to join UCAN and LatAm. This is a central piece of extracting optimal value, as it enables more of a user-by-user means of advertising. No more spray and pray.

Misc. Notes:

Great quarter for a company that will be seen as a safe haven for as long as trade war drama persists. I think they would’ve raised guidance in normal times, and I expect outperformance through 2025. Whether it’s price hikes, ad proliferation, content diversification, globalization, margin expansion or anything else you want to look at… they are killing it. See the last few years of margin-accretive compounding for evidence. There’s nothing negative to pick at here.