TSMC has raised its full-year revenue growth forecast to 44%, upping its outlook as demand for advanced-node chips (the smallest, most powerful transistor designs) keeps climbing on the back of the AI buildout.

The revision lands as AI infrastructure spending shows no sign of cooling. TSMC, the Taiwanese contract manufacturer that produces silicon for nearly every major AI accelerator on the market, now expects nearly half again as much revenue this year as it booked last year — a pace few companies of its scale ever reach.

Advanced nodes did the heavy lifting. That's the technical term for leading-edge process geometries, the cutting-edge manufacturing tiers where AI processors and high-end compute silicon are made. Customers have been stacking orders there faster than TSMC can add capacity, and the updated 44% figure reflects that pressure flowing straight into the top line.

What does the forecast change?

For TSMC's customers — the chip designers who depend on its fabs — the raised outlook signals that demand is strong enough to keep justifying aggressive capacity investment. For investors, it confirms the AI boom is still deepening rather than plateauing. And for competitors, it's a reminder of the gap: advanced-node demand keeps consolidating around the one foundry that can supply it at volume.

The forecast raise also functions as a demand gauge for the whole electronics supply chain. When the dominant foundry lifts its growth projection by this magnitude, equipment suppliers, packaging partners and materials vendors downstream typically feel it within quarters.

Why the AI boom keeps widening

AI accelerators are large, expensive chips. Each one consumes more wafer area than a typical smartphone processor, which means every additional AI datacenter deployment translates into outsized foundry revenue. TSMC sits at the choke point of that flow: it manufactures the overwhelming majority of high-end AI compute silicon, so rising AI spending reaches its books quickly and at scale.

Advanced-node demand climbing is the mechanism behind the number. AI chips require the tightest transistor densities available, and that requirement concentrates orders on the newest process tiers — precisely the capacity TSMC has spent years and billions expanding.

The 44% forecast doesn't describe a rebound. It describes acceleration. And it arrives after a period in which the foundry had already been steadily upgrading its expectations as AI orders stacked up.

What it means in one line: the world's most important chipmaker now believes its AI-driven growth is running hotter than it previously told the market — and it has put a hard number on it.