Samsung published a deep dive on its 4nm FinFET process through its official semiconductor blog. That’s an unusual move for a company that usually lets its fab roadmaps do the talking.

Pay attention: Samsung began mass-producing 4nm in 2021. Six years of production data doesn’t sound glamorous, but in the foundry business, accumulated process knowledge is currency.

According to Business Korea, Samsung says it has used that data to drive meaningful improvements in interconnect resistance and capacitance delay, cutting RC delay by around 26% compared to its previous generation.

Within the same 4nm node, Samsung can produce performance-hungry chips for AI accelerators and low-power designs for automotive or mobile. Different threshold-voltage options let customers tune to their specific constraints.

Samsung is applying its 4nm process to the base die of HBM4. The 4nm process handles cramming enormous data throughput into a confined physical space by keeping power loss down while pushing integration density up.

The Korean tech giant is running a two-track foundry strategy. It’s chasing the bleeding edge with its 2nm gate-all-around process, and it’s leaning into 4nm as a mature, stable option for customers who can’t afford yield uncertainty.

Meanwhile, the company has historically struggled to communicate the value of what it already has. It chases announcements about future process nodes while underplaying production capabilities that are actually ready to go.

This blog post, understated as it is, reads like Samsung trying to correct that habit.

The post Samsung’s 4nm FinFET process is central to its HBM4 strategy appeared first on Sammy Fans.