Gate-All-Around

If wrapping a transistor’s gate around three sides of its channel — the FinFET approach — was such a big improvement, the obvious next question is: why not wrap it around all four? That’s exactly what Gate-All-Around (GAA) transistors do, and it’s the structure the leading edge of the chip industry has been transitioning to since around the 3nm process generation. It’s less a wild departure from the FinFET and more its logical, more complete successor.

Finishing what the FinFET started

Recall that a FinFET stands its channel up as a vertical fin and wraps the gate around the top and two sidewalls — three out of four sides. The bottom of the fin, however, stays physically attached to the substrate below it, so the gate never gets full control over that fourth face. As transistors kept shrinking even further, that last uncontrolled side started to matter again: the same leakage and control problems that motivated the shift to FinFETs in the first place began creeping back in.

The fix is architecturally elegant: instead of a fin standing on the substrate, the channel is reshaped into one or more thin horizontal sheets (or, in some designs, thin wires) that are physically separated from the substrate and from each other, with gate material filling in all the way around each one — top, bottom, and both sides.

🧩 Think of it like… the difference between pinching a hose with your fingers (FinFET — three sides squeezed) versus closing your entire fist all the way around it (GAA — all four sides squeezed, nothing left uncontrolled). A closed fist gives you the most complete shutoff possible, because there's no gap left for anything to sneak through.

Nanosheets: stacking for extra current

A single fully-wrapped sheet would actually carry less current than a fin, because a flat sheet has less surface area facing the gate than a tall vertical fin does. The industry’s solution is to stack several thin nanosheets on top of each other, one above another, with the gate material filling the gaps between them as well as wrapping around the outside of the whole stack. Each stacked sheet acts as its own fully-gated channel, and stacking two or three of them recovers — and often exceeds — the current-carrying capacity of a FinFET, while keeping the superior all-around electrostatic control.

The thin dark outline around each pink nanosheet in the diagram represents the gate oxide and gate metal layer that completely encases the channel — notice there’s no uncontrolled face left anywhere, unlike the FinFET’s exposed base.

Why this matters for scaling

Full electrostatic control has a direct, practical payoff: it lets engineers shrink the channel length further without the transistor’s off-state leakage getting out of hand, and it allows some extra design flexibility that FinFETs don’t offer. Because each nanosheet’s width can be tuned independently (unlike a FinFET’s height, which is more rigidly constrained by the fin etching process), GAA designs give chip architects a new knob — sheet width — to trade off switching speed against power consumption on a per-transistor basis. That flexibility is part of why GAA is expected to remain the leading transistor architecture for at least the next several process generations, even as node names continue to be more marketing label than physical measurement (see What Does “7nm” Actually Mean?).

Different companies use different brand names for their version of gate-all-around technology — for example, “MBCFET” and “nanosheet transistors” both refer to variations on the same underlying idea: full, all-sides gate control over one or more stacked channel sheets.

Manufacturing complexity

None of this comes free. Building a GAA transistor requires precisely growing alternating layers of two different materials, then selectively removing one of them to leave the nanosheets suspended with empty gaps around them, and finally depositing gate material into those tiny gaps so it wraps completely around each sheet — a much more intricate sequence of etching and deposition steps than a FinFET requires. It’s a clear example of a pattern that repeats throughout chip manufacturing history: each step toward better transistor performance tends to demand a proportionally larger jump in manufacturing sophistication.

Key takeaways

Gate-All-Around (GAA) transistors wrap the gate completely around the channel on all four sides, closing the one gap FinFETs leave at the base of the fin.
To keep current-carrying capacity high despite using flatter channels, GAA designs stack several thin nanosheets, each fully gated, on top of one another.
Full electrostatic control means less leakage and more room to keep shrinking channel length without losing switching behavior.
Independently tunable nanosheet width gives chip designers a new lever for balancing speed against power, which FinFETs didn’t offer as directly.
GAA transistors require substantially more complex manufacturing steps than FinFETs, continuing chipmaking’s long trend of rising process complexity in exchange for better performance.

FinFET EUV Lithography