Explain It in 30 Seconds
As chips become denser, power delivery competes with signal routing for limited space on the front side of the chip.
A16 moves key power-delivery structures to the back side.
The basic idea is:
front side
more room for signals
back side
power deliveryThat can reduce power-delivery losses and free routing resources for logic.
Where A16 Sits in TSMC
A16 is part of TSMC's leading-edge logic roadmap.
It builds on the nanosheet era introduced with N2 but adds a new power-delivery architecture.
The Problem A16 Solves
High-performance chips do not only need more transistors.
They need power delivered efficiently across very dense logic.
As routing becomes more crowded:
signal wires
That can create:
- voltage drop;
- routing pressure;
- density constraints;
- performance limits.
A16 addresses that system-level problem through backside power delivery. [S9]
How It Works — Simple Version
Traditional chip layouts deliver most power from the same side used for much of the signal routing.
A16's Super Power Rail moves power delivery to the back side of the wafer.
TSMC says this frees front-side routing resources for signals and reduces IR drop. [S9]
The high-level mechanism is:
N2P vs. A16
TSMC positions A16 for cases where backside power delivery creates enough value to justify the different design approach.
Compared with N2P, TSMC targets A16 for:
- 8–10% speed improvement at the same voltage;
- 15–20% power reduction at the same speed;
- up to 1.10× chip density. [S9]
These are TSMC's stated technology targets, not universal guarantees for every customer product.
Who Uses A16?
The strongest fit is high-performance computing.
Examples of product types that may benefit include:
- AI accelerators;
- data-center processors;
- large compute ASICs;
- designs with dense power-delivery networks.
TSMC explicitly describes A16 as best suited to HPC products with complex signal routes and dense power networks. [S9]
What A16 Depends On
A16 depends on:
- nanosheet transistor manufacturing;
- backside wafer processing;
- advanced lithography;
- new design rules;
- EDA and IP support;
- customer willingness to adopt the architecture;
- packaging suitable for high-performance products.
Why A16 Matters to TSMC
A16 shows that the foundry roadmap is changing in two dimensions at once.
Traditional scaling focuses on transistor geometry.
A16 also changes where power is delivered.
That means future foundry competition is increasingly about system-level engineering, not only smaller transistors.
What Most People Misunderstand
"A16 is simply a smaller number after N2."
No.
TSMC uses the A16 name for a technology offering that adds a distinct backside power-delivery architecture.
"Backside power means power is generated on the back of the chip."
No.
It describes where power-routing structures are delivered, not where electrical energy originates.
"Every N2 customer should automatically move to A16."
No.
The value depends on the customer's product architecture and routing/power needs.
Common Questions
Is A16 in production?
Does A16 use nanosheet transistors?
Yes.
Is A16 mainly for smartphones?
TSMC emphasizes HPC as the strongest fit because of complex routing and power-delivery requirements.