TSMC posted a 77% jump in second-quarter profit to T$706.6 billion on Thursday, and the company tied the record result to surging demand for AI chips. For customers such as Nvidia and Apple, the report says orders for advanced manufacturing remain the key issue as TSMC heads into its guidance call.
Wen-Yee Lee, Ben Blanchard and Faith Hung reported that net profit for April to June reached T$706.6 billion, or $21.99 billion, above an LSEG SmartEstimate of T$632.6 billion compiled from 18 analysts. TSMC said the increase came from demand for advanced chips used in artificial intelligence applications, while analysts said demand for its 3-nanometre and 2-nanometre process technologies and CoWoS packaging remained strong.
TSMC revenue rose 36%
36% was the rise TSMC announced in second-quarter revenue on Monday, a separate sign that the company’s AI-linked business remained firm before the profit release. Taipei-listed shares have gained 59% so far this year, reflecting how closely the stock has tracked the pace of demand for advanced chips and packaging.
TSMC stock rises as C.C. Wei signals $56 billion capex ceiling also sits against the same backdrop: investors are weighing whether the company keeps spending near the high end of its earlier plan, after it said in April that 2026 capital expenditure would be at the high end of $52 billion to $56 billion.
TSMC guidance call at 0600 GMT
0600 GMT is when TSMC plans to hold its earnings call and give third-quarter and full-year guidance. That call is the next point where the company can turn a profit beat into a clearer spending and revenue path, especially after analysts had expected it to raise its outlook.
$165 billion is the scale of TSMC’s planned chip-factory investment in Arizona, a reminder that the company is still committing heavily while demand stays concentrated in AI-related products. If the call leaves the outlook unchanged, the profit surprise still stands as evidence of strong execution, but it would also leave open whether management thinks the current AI cycle can support a bigger full-year plan.
