$Taiwan Semiconductor Manufacturing(TSM)$ delivered another blockbuster quarter. Yet TSMC and $NVIDIA (NVDA)$ fell as investors continued selling semiconductor stocks.
If AI demand is so strong, why isn’t the market celebrating?
📊 1. The numbers are impressive
TSMC’s third-quarter 2026 results showed just how powerful the AI chip boom has become:
🔥 Revenue reached NT$1.494 trillion, up 50.9% year over year.
🔥 Net profit hit NT$763.1 billion, up 68.7%.
🔥 Operating margin reached 58.7%.
🔥 Fourth-quarter revenue guidance came in at US$52.3–54.3 billion, above the previous quarter’s level.
These aren’t the numbers of a business experiencing an obvious demand collapse. TSMC continues to benefit from demand for advanced chips used in AI infrastructure. 
So why the selling?
💰 2. The market is looking beyond today’s earnings
Here’s the problem: investors are no longer asking only how much AI companies can sell. They’re asking how much they must spend to keep growing.
TSMC plans US$60–64 billion in capital expenditure for 2026, with spending expected to rise significantly over 2026–2028 as it expands capacity.
That investment makes sense if AI demand continues growing. But new factories, advanced manufacturing equipment and supporting infrastructure require enormous upfront spending.
Invest too little, and TSMC risks losing opportunities to competitors. Invest too much, and it risks building expensive capacity that takes longer than expected to generate an adequate return.
This is the balancing act facing the entire AI supply chain. 
⚠️ 3. Nvidia faces a different question
TSMC manufactures chips. Nvidia sells the computing platforms that power much of the AI boom.
Both benefit from growing AI demand, but their investment risks are different.
For Nvidia, the key question is whether customers will continue spending heavily on AI infrastructure and generate sufficient returns to justify further investment.
For TSMC, the question is whether it can convert rising chip demand into sustainable profits while managing its expanding capital requirements.
Strong orders today are encouraging. However, investors also need confidence that demand will remain strong enough to justify tomorrow’s spending.
📉 4. Why good earnings can still mean falling share prices
Markets price in expectations, not just reported results.
When investors already expect exceptional growth, even record profits may fail to push a share price higher if the outlook raises new concerns.
Recent selling across semiconductor stocks has also coincided with hotter-than-expected inflation and rising interest-rate concerns. Higher financing costs can put pressure on companies investing billions in AI infrastructure. 
This doesn’t automatically mean the AI boom is ending.
It may mean the market is becoming more selective about valuations, financing costs and the returns expected from AI spending.
🎯 My takeaway
I wouldn’t interpret one day’s share-price decline as proof that AI demand is weakening.
TSMC’s results demonstrate that demand for advanced chips remains substantial. But the next stage of the AI trade may depend less on impressive revenue growth alone and more on whether companies can turn massive investment into sustainable cash flow.
The winners may not simply be the companies spending the most on AI, but those that generate the strongest returns from that spending.
💬 If you had to choose today, would you favour TSMC, which manufactures the chips, or Nvidia, which sells the AI computing platforms? And what would make you confident that their current valuations are justified?
Sharing this perspective for discussion purposes only.
Comments