How Autonomous AI Is Transforming Chip and System Design

Watch on YouTube

Now Playing

Transcript

57 segments

0:02

The race to build AI factories is on.

0:05

But each new generation of chips powering

0:08

them is more complex than the last.

0:11

Every chip travels the same path—

0:14

design,

0:15

verification,

0:16

implementation,

0:17

signoff.

0:18

The systems they power

0:19

must meet relentless demands:

0:21

performance, power, area, thermal.

0:25

Hundreds of engineers.

0:26

Months of work.

0:27

One goal:

0:28

first-pass silicon.

0:30

The answer:

0:31

autonomous

0:32

AI agents

0:33

across the entire EDA

0:34

and systems engineering stack.

0:37

At design,

0:38

Cadence's autonomous AI engineer

0:40

compresses RTL development

0:42

from weeks to hours,

0:43

and drives the flow itself.

0:45

In verification and implementation,

0:48

Synopsys's broad set of agents

0:50

autonomously handle RTL to GDS,

0:54

with human supervision.

0:55

And Synopsys PrimeSim SPICE

0:57

accelerates simulation up to

0:59

18 times faster on NVIDIA GPUs.

1:03

At signoff, Siemens Fuse EDA AI Agent

1:06

performs autonomous physical verification.

1:09

Next, Cadence AuraStack AI

1:12

unifies 3D-IC packaging

1:14

and PCB design from exploration

1:17

through physical implementation

1:19

and signoff, while NVIDIA cuDSS

1:22

accelerates thermal closure

1:24

for stacked systems.

1:26

Beyond the chip, agents

1:27

use NVIDIA PhysicsNeMo

1:29

to predict fluid,

1:30

thermal, and structural behavior—

1:33

while Synopsys extends from EDA

1:36

into System Design,

1:37

silicon to systems,

1:39

accelerated by NVIDIA cuISS.

1:43

Fewer iterations

1:45

give engineers more time to invent

1:47

what comes next.

1:48

This is how

1:49

next-generation chips

1:50

get built with NVIDIA technology,

1:53

designed, verified, and delivered

1:55

at the speed AI and the world demand.

Interactive Summary

Ask follow-up questions or revisit key timestamps.

The video outlines how the chip design industry is transitioning to autonomous AI agents to manage the increasing complexity of modern semiconductor manufacturing. By integrating AI across the entire EDA stack—from design and verification to signoff and system packaging—companies like Cadence, Synopsys, and Siemens are using NVIDIA technology to drastically accelerate development timelines, reduce iterations, and achieve first-pass silicon.