TSMC Profit Jumps 77%, ASML Raises Its Outlook: How Much Further Can the AI Hardware Expansion Run?
$Taiwan Semiconductor Manufacturing(TSM)$ $ASML Holding NV(ASML)$
TSMC and ASML have delivered another strong signal for the AI semiconductor cycle.
TSMC reported second-quarter revenue of $40.2 billion, a gross margin of 67.7%, and net profit growth of 77.4% year over year.
ASML, meanwhile, raised its 2026 revenue outlook from €36 billion–€40 billion to €43 billion–€45 billion, while preparing to expand EUV and immersion DUV capacity again in 2027 and 2028.
Together, the two earnings reports confirm that demand for AI chips, advanced process nodes, HBM and semiconductor equipment remains strong.
The next question is moving downstream:
Can Microsoft, Alphabet, Meta and Amazon turn all of this infrastructure spending into enough cloud revenue, advertising growth, subscriptions and AI-product sales?
TSMC: AI demand is already turning into profit
TSMC’s second-quarter revenue reached $40.2 billion, up 33.7% from a year earlier and 12% sequentially, landing near the upper end of its guidance.
Gross margin rose to 67.7%, while operating margin reached 60.3%, both above the company’s previous forecast range.
Net profit climbed to NT$706.56 billion, setting a new record.
The reported profit growth included roughly NT$63.2 billion in gains related to the disposal and revaluation of Vanguard International Semiconductor shares.
Even after excluding that item, the underlying earnings performance remained stronger than expected.
The more important signals came from margins and utilization.
TSMC maintained a gross margin close to 68% despite overseas fab expansion, rising depreciation and the early ramp of 2-nanometer production.
That suggests advanced-node utilization, product mix and pricing power remain exceptionally strong.
TSMC is increasingly becoming an AI infrastructure company
High-performance computing accounted for 66% of total revenue and grew 20% sequentially.
Smartphone revenue fell 4% quarter over quarter and represented 22% of total sales.
TSMC’s growth engine is becoming increasingly tied to AI accelerators, data-center computing and advanced process technology.
The process-node mix tells the same story:
-
2nm contributed 3% of wafer revenue for the first time
-
3nm contributed 30%
-
5nm contributed 33%
-
7nm and below accounted for 77% in total
The early contribution from 2nm and continued strength in 3nm indicate that advanced-node demand has not entered a clear slowdown.
TSMC expects third-quarter revenue of $44.6 billion–$45.8 billion, with a midpoint of roughly $45.2 billion.
That would represent another sequential increase of about 12%.
Gross margin is expected to moderate to 65%–67%, while operating margin is guided to 56%–58%.
The slight margin decline likely reflects the initial costs of 2nm production, overseas-fab depreciation and new-capacity ramp expenses.
ASML: the outlook matters more than the quarter
ASML reported second-quarter revenue of €9.3 billion, a gross margin of 54%, net profit of €2.9 billion and earnings per share of €7.59.
The quarter was helped by stronger-than-expected revenue from installed-base management, including upgrades, maintenance and productivity improvements for existing equipment.
As the global EUV installed base grows, software upgrades and service revenue are becoming a larger, recurring and high-margin part of ASML’s business.
But the biggest surprise came from guidance.
ASML raised its 2026 revenue outlook to €43 billion–€45 billion.
Third-quarter revenue is expected to reach €11 billion–€12 billion, implying growth of roughly 23% from the second-quarter midpoint.
Gross margin is projected at 55%–57%.
Management said customers are increasing capital spending and accelerating capacity additions for advanced logic and DRAM.
Memory expansion may be one of the strongest signals
ASML expects 2026 revenue from advanced logic systems to grow by more than 25%.
Memory-related system revenue is expected to grow by more than 75%.
EUV revenue is expected to rise by more than 45%, while installed-base management revenue may grow by more than 30%.
The projected 75% increase in memory equipment revenue is especially important.
ASML said higher DDR and HBM prices, tight supply and stronger demand are pushing memory manufacturers to expand production.
More advanced DRAM nodes also require greater use of EUV and immersion DUV equipment.
This suggests that the current memory capex cycle is being driven primarily by DRAM and HBM.
For U.S. investors, $Micron Technology(MU)$ remains one of the clearest listed beneficiaries.
$SanDisk(SNDK)$ is more directly exposed to NAND, so its benefit is more indirect and still depends on NAND pricing and enterprise flash demand.
Equipment demand is now visible into 2028
ASML plans to increase Low-NA EUV capacity by roughly 30% in 2027.
The additional capacity is already close to fully covered by orders.
The company has also received significant demand indications for 2028 and is evaluating another capacity increase of roughly 30%.
Immersion DUV capacity is also expected to expand by 30% in 2027, with another increase possible in 2028.
This matters because TSMC is confirming current wafer demand, while ASML is confirming future equipment demand.
When chip manufacturers are willing to reserve equipment years in advance, it suggests the AI-chip and memory expansion plans still have relatively high visibility.
Why does stronger upstream demand make investors more nervous?
TSMC and ASML have confirmed that AI-chip orders are real and semiconductor capacity is still expanding.
But the larger the buildout becomes, the more investors will question the final return on that capital.
The chain works roughly like this:
ASML supplies semiconductor equipment.
TSMC and memory manufacturers build capacity.
Nvidia, AMD and Broadcom design chips.
Microsoft, Alphabet, Meta and Amazon buy those chips and construct data centers.
The hyperscalers then need to generate returns through cloud services, advertising, enterprise software, AI agents and subscriptions.
The upstream part of the chain has already demonstrated strong demand.
The next major test comes from the companies ultimately paying for the infrastructure.
In the upcoming hyperscaler earnings reports, investors should focus on four areas:
-
Is cloud growth continuing to accelerate?
-
Are AI revenue and contracted backlogs increasing?
-
Are capital-expenditure forecasts still rising?
-
Are free cash flow and margins being materially pressured?
The ideal combination would be accelerating AI revenue, strong cloud growth and sustained capital spending.
The more concerning combination would be continued capex expansion alongside slowing cloud growth, weak AI monetization and falling free cash flow.
Alphabet is scheduled to report on July 22, while Microsoft and Meta are expected to report on July 29.
Those results could become the next major test for the AI hardware rally.
Where are we in the semiconductor cycle?
Based on the TSMC and ASML reports, the AI semiconductor cycle still appears to be in the middle of a capacity-expansion phase.
TSMC expects another double-digit sequential revenue increase in the third quarter.
ASML’s 2027 EUV capacity is already close to being covered by orders.
Advanced logic, HBM and DRAM are all driving additional equipment demand.
These signals suggest that industry fundamentals remain supported over the next several quarters.
The bigger short-term risks are concentrated in market positioning and valuation:
-
Semiconductor valuations remain elevated
-
Some AI trades are crowded
-
Korean chip stocks have recently experienced deleveraging
-
Investors are increasingly questioning hyperscaler returns on capital
Strong earnings do not guarantee rising share prices.
The more serious warning signs would include hyperscalers cutting capex, TSMC customers reducing advanced-node orders, ASML order visibility and delivery times shrinking, or HBM pricing and memory capex weakening at the same time.
Those signals have not yet appeared in a broad and coordinated way.
Stocks to watch
Foundry:
$Taiwan Semiconductor(TSM)$
Semiconductor equipment:
$ASML Holding(ASML)$
$Applied Materials(AMAT)$
$Lam Research(LRCX)$
$KLA Corporation(KLAC)$
AI chips:
$NVIDIA(NVDA)$
$Advanced Micro Devices(AMD)$
$Broadcom(AVGO)$
Memory expansion:
$Micron Technology(MU)$
$SanDisk(SNDK)$
Advanced packaging:
$Amkor Technology(AMKR)$
$ASMPT(00522.HK)$
TigerComments Take
TSMC and ASML delivered the same message:
The AI hardware expansion has not hit the brakes.
But the next phase of the trade will increasingly be determined by Microsoft, Alphabet, Meta and Amazon.
The upstream supply chain has proven that the orders exist.
Now the downstream buyers need to prove those orders can turn into AI revenue, profit and cash flow.
Where do you think the AI semiconductor cycle goes next?
A. Hyperscaler earnings confirm demand, and chip stocks make new highs after the correction
B. Fundamentals remain strong, but valuations need more time to digest
C. Capex is growing too quickly, creating future oversupply risk
D. Focus only on the core leaders: TSM / ASML / NVDA
Do you worry more about missing the next leg of the AI hardware rally, or about too much capacity being built for the future?
Disclaimer: This post is for market discussion only and does not constitute investment advice. Investing carries risk.
Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.

For me, the next catalyst is no longer semiconductor earnings but the hyperscalers. Microsoft, Meta, Alphabet and Amazon need to prove AI spending can drive faster cloud growth, stronger AI revenue and healthy free cash flow. If they deliver, I believe the AI cycle still has room to run.
I'm not too worried about missing the rally or an imminent oversupply. I'm staying focused on leaders like TSMC, ASML and NVIDIA while watching Micron closely. As long as orders remain strong and hyperscalers keep investing, this still feels like the middle of the AI investment cycle rather than the end.
@TigerClub @TigerStars @Tiger_comments