
TSMC and ASML both report midweek, and those two prints open the Q2 earnings season for the entire chip complex. TSM trades around $438.09 going in, which sounds calm until you look at what the rest of the sector has done in the past few weeks. Intel is down roughly 4.6% today and has slid hard since flagging a yield delay on its 18A process, a Samsung-earnings miss dragged the whole group lower, and Bank of America put the words "bubble risk" next to AI chip valuations in a research note that traders are still arguing about.
Chip Sector Snapshot, July 13, 2026
- TSM price: around $438.09
- Earnings window: both companies report midweek, with ASML delivering its print first and TSMC following roughly a day later
- What to watch at ASML: net bookings and the order backlog, the leading indicator for future fab capacity
- What to watch at TSMC: revenue guidance and the advanced-node mix, the read on actual AI chip demand
- Sector backdrop: INTC down about 4.6% today, a Samsung-triggered selloff still in the tape, a Bank of Americabubble-risk warning, and Micron raising its US investment commitment to $250 billion-plus through 2035
The market is split down the middle between traders who think AI demand is structural and traders who think the valuations ran too far ahead of the revenue. Here is what each company actually reports, the specific lines that decide how the sector trades, and what a beat or a miss would do to everything else in the AI chip complex.
What ASML Reports and Why Bookings Are the Leading Indicator
ASML is the only company on earth that builds extreme ultraviolet lithography machines, the systems that print the finest features on an advanced chip. Every leading-edge fab in the world buys them from one Dutch supplier, which makes ASML the closest thing the industry has to a single point of truth about future capacity.
The number that moves the stock is not revenue. It is net bookings, the value of new machine orders taken during the quarter. Revenue recognized this quarter reflects orders placed one to three years ago, so it describes the past, while bookings tell you what the foundries and memory makers are committing to build in 2027 and 2028.
Think of it as the difference between a restaurant's receipts and its reservation book. The receipts tell you last night went fine. The reservation book tells you if anybody is coming back.
A strong bookings print says the fabs are still racing to add leading-edge capacity for AI accelerators and high-bandwidth memory. A soft print, even alongside good revenue, says customers are pausing, and that is the single most bearish datapoint the sector could produce this week. ASML posts the figures on its ASML investor financial results page, where the bookings line sits at the top of the release.
What TSMC Reports and Why the Guidance Line Decides the Week
TSMC manufactures the chips that ASML's machines make possible, and it does so for almost everyone that matters in AI. The silicon behind the accelerators filling data centers is fabricated on TSMC's most advanced nodes, which makes its order book the cleanest proxy available for real AI hardware demand rather than announced AI hardware demand.
Three lines carry the print. Forward revenue guidance is the headline, because it is TSMC's own statement about what customers have committed to buy next quarter. The advanced-node mix, meaning the share of wafer revenue coming from 3nm and 5nm production, shows how much of the business is genuinely AI and high-performance computing rather than mature-node phone and automotive silicon. And full-year capex tells you how much TSMC will spend to serve the demand it says it sees, which loops straight back into ASML's bookings.
Gross margin sits underneath all of it, because TSMC has pricing power no other foundry can match right now. If margins hold while capex rises, the bull case gets a large piece of evidence. The numbers land on the TSMC quarterly results page, and the call commentary usually matters more than the reported quarter. TSMC guidance also sets expectations for NVIDIA and the rest of the AI chip designers that depend on its capacity, because a designer cannot ship what its foundry cannot fabricate.
The Numbers to Watch in Both Prints
Two companies, two very different signals. This is the cheat sheet.
|
Metric
|
Company
|
What it tells you
|
The bearish read
|
|
Net bookings
|
ASML
|
Fab capacity being committed for 2027 and 2028
|
A sequential drop means customers are pausing the capex cycle
|
|
Order backlog
|
ASML
|
How much visibility the equipment cycle still has
|
A shrinking backlog removes the safety net under the whole complex
|
|
Forward revenue guidance
|
TSMC
|
Real committed AI chip demand for next quarter
|
Guidance below the prior trend says the order book is cooling
|
|
Advanced-node mix
|
TSMC
|
How much revenue is genuinely AI and HPC
|
A flat or falling 3nm share means the AI mix is not deepening
|
|
Full-year capex
|
TSMC
|
Confidence in demand two years out
|
A capex cut is the loudest exhaustion signal available
|
|
Gross margin
|
TSMC
|
Pricing power at the leading edge
|
Margin compression means the foundry is discounting to fill capacity
|
Read them in that order. ASML frames the ceiling on future capacity, TSMC tells you how fast the current capacity is being consumed, and the market prices the gap between the two.
The Bull Case for the AI Chip Trade
The bulls do not need either company to beat by a wide margin. They need the direction of travel to hold, and several things going into this week suggest it will.
Micron pushed its US investment commitment past $250 billion through 2035, and companies do not commit capital on that timeline against a demand curve they expect to roll over next year. AI memory demand has been running hot enough that high-bandwidth memory became the bottleneck rather than the logic die, which is the kind of shortage that keeps foundry utilization high. Our breakdown of why Micron sits at the center of the AI memory build-out covers how tightly that supply chain is wound.
The structural argument is simple. Hyperscaler capex budgets are set annually and have been revised up rather than down, the chips those budgets buy get fabricated on TSMC's advanced nodes, and the fabs that make them get built with ASML's machines. Until one of those links visibly cracks, the bull case reads every selloff in the group as positioning noise rather than a demand signal.
And the sector's recent weakness has company-specific explanations. Intel's slide came from a yield problem on its own process rather than from customers cutting orders, which the Intel press release archive makes clear enough. A Samsung miss is a Samsung problem before it is an industry problem, as our read on Samsung's 2026 setup lays out.
The Bear Case and Why Bank of America Flagged Bubble Risk
The bear case does not argue that AI demand is fake. It argues that the price already assumes the demand continues for years, and that leaves no margin for a single soft quarter.
Bank of America's bubble-risk language landed because AI chip multiples have been expanding faster than the earnings underneath them, which is the textbook setup for a violent repricing on any disappointment. A stock priced for perfection does not need bad news to fall. It only needs news that is merely fine, and that is a low bar for two companies reporting in the same 48-hour window.
The tape is already flashing warnings. Intel is down about 4.6% today on the 18A yield delay, and the market read that as evidence that leading-edge manufacturing is harder and more expensive than the roadmaps imply. The Samsung-triggered selloff then hit names that had nothing to do with Samsung, which is what happens when positioning is crowded and everyone reaches for the same exit. Broader coverage of those swings sits in CNBC's semiconductor section, and the pattern has been consistent for weeks.
The specific bear trigger this week is an ASML bookings miss paired with cautious TSMC guidance. That combination would say the equipment cycle is cooling at the same time the foundry sees demand flattening, and there is no comfortable way to read those two together. The rerating in ARM earlier this year is a reminder of how fast an AI-adjacent premium comes out of a stock when the story wobbles.
How to Trade the Print Without Guessing the Direction
Earnings are binary events, and the honest answer is that nobody knows which way two prints in one week resolve. This is where retail consistently loses money, sizing a full position into a coin flip and then getting stopped out on the initial spike before the real move develops.
The setup that respects the uncertainty defines the levels first. TSM around $438 is the reference price, and the reaction matters more than the headline number. A gap higher that holds through the following session says institutions are adding rather than selling relief, while a gap higher that fades back through the pre-print level within a day is distribution.
Watch the sequencing, too. ASML reports first, so its bookings number becomes the lens the market uses to read TSMC a day later. Strong bookings followed by strong guidance is the clean bull confirmation, and strong bookings followed by weak guidance is the ugliest outcome, because it means machines are being ordered into demand that is already softening. Traders comparing the AI semiconductor names against each other will find that the divergence between equipment and foundry is where the sector's real story gets told. Live quotes and the reaction to each release are on the TSM quote page at Yahoo Finance, but no data feed saves a position that was too large going in.
Frequently Asked Questions
When does TSMC report earnings?
TSMC reports midweek, following ASML by roughly a day, and both prints land in the same short window. The exact timing is posted on TSMC's investor relations site, and the earnings call commentary typically moves the stock more than the reported quarter itself.
What are ASML bookings and why do traders watch them?
Bookings are the value of new lithography machine orders taken during the quarter, and they are the industry's leading indicator because those machines get delivered one to three years later. Revenue tells you what fabs ordered in the past, while bookings tell you what they are committing to build in the future, which is the number that prices the AI capex cycle.
Is TSM stock a buy before earnings?
Buying directly into a binary event is a bet on the reaction, not on the business. The more disciplined approach is to size small into the print or wait for the first full session after guidance, when the market has told you how it plans to interpret the numbers.
What happens to AI chip stocks if TSMC misses?
A guidance miss at TSMC would hit the chip designers that depend on its capacity, because a foundry that expects less business is a foundry whose customers ordered less silicon. Expect the drawdown to spread beyond the direct AI names into equipment and memory, since the entire complex trades off the same demand assumption.
The Bottom Line
This week is the first real test of the AI chip trade since Bank of America put bubble risk on the table, and it arrives with the sector already jumpy. Strong ASML bookings plus TSMC guidance above trend confirms the capex cycle has legs, and the recent weakness in INTC and the Samsung-driven selloff get filed as noise. Soft bookings plus cautious guidance is the exhaustion signal the bears have been waiting for, and it would reprice every AI-adjacent name in the complex, not only the two reporting. The awkward middle outcome, strong bookings with weak guidance, is the one nobody is positioned for and the one worth watching most closely. TSM around $438 is the line in the sand. Hold it after both prints and the trade is intact. Lose it on volume and the market has answered the question it has been asking all month.
This article is for informational purposes only and does not constitute financial or investment advice. Cryptocurrency trading involves substantial risk. Always conduct your own research before making trading decisions.
