TSMC has posted its highest quarterly profit on record, the company reported, driven by surging demand for chips used in artificial intelligence applications.

The third-quarter result marks a new peak for the world's largest contract chipmaker and signals that spending on AI infrastructure continues to translate directly into silicon orders. Alongside the earnings announcement, TSMC confirmed it is considering building a fabrication plant in Texas.

What does the record profit signal?

The profit surge confirms TSMC as a primary beneficiary of the AI build-out. Customers designing AI systems rely on the company's advanced process nodes, and demand for that capacity pushed quarterly earnings past every prior quarter in the firm's history.

For the wider semiconductor supply chain, the result functions as a demand gauge. When TSMC's order books tighten, equipment vendors, materials suppliers and packaging partners typically follow. A record quarter at the top of the chain points to sustained upstream pressure on capacity through the coming quarters.

The takeaway: AI chip demand has converted into record earnings at the foundry level, not just at the design level.

What would a Texas fab change?

TSMC said it is weighing a fabrication facility in Texas. The company has not yet committed to the project, and it disclosed no timeline, investment figure or capacity target alongside the statement.

A Texas site would add to TSMC's existing footprint in the United States. For US customers — particularly those designing AI processors — an additional domestic fab would shorten supply lines and align with federal policy favoring onshore semiconductor manufacturing.

The company framed the Texas option as under evaluation. Any final decision would follow TSMC's standard practice of announcing sites with committed budgets and construction schedules.

The takeaway: Texas is on the table, but TSMC has made no commitment beyond confirming the evaluation.

Why the AI demand matters now

The record quarter lands amid a broad expansion of AI datacenter construction. Every new training and inference cluster requires advanced logic chips, and TSMC fabricates the majority of leading-edge designs that power them.

That positions the foundry as a chokepoint and a beneficiary at once: capacity constraints limit how fast AI hardware ships, while the same constraints support pricing and utilization rates at TSMC's plants.

The takeaway: as long as AI clusters keep expanding, TSMC's utilization and margins face upward pressure.

TSMC's next disclosures — a Texas decision and full-year guidance — will show whether the record quarter was a peak or a baseline.