Scaling Custom Silicon in China: Technology Is Only the Beginning

But while we're talking about complementary, let's look at the focus of both companies, right? Both Bayer and Adorsys seem to be focused on data center, HPC, edge AI, and automotive solutions, right? So why these markets and how do you see them evolving over the next few years? And let's start again with you, Keng Su.

Okay.

Basically, these are all AI compute.

So they are experiencing massive growth.

Compute is moving from a general purpose processes towards application specific AI centric SOC.

And customers increasingly need custom silicon.

So looking at the worldwide AI infrastructure spending for the past three years and the forecast for the next five years, we know that the massive growth is going to continue.

Therefore, Adolphe's pursuing AI compute is a strategically sensible direction.

That makes sense.

And I think these are also complex designs that require specialized solutions.

And maybe, Shailesh, you can give a bit more color because you have seen so many of the customers in the market look for differentiated solutions.

Yeah, absolutely.

So certainly from a BIAS standpoint, we are seeing lots of opportunities in the AI and infrastructure space.

I will say largely because this is where the biggest disruption is happening.

The compute architecture is fundamentally being reinvented.

I remember when I was at Intel in 2018 onwards, Xeon had a 99% market share in data center compute.

The entire data center stack was built around CPUs and that to Intel CPUs.

Now, if you look at data center today, it's fundamentally different.

Everything is built around GPUs, in-house accelerators, highly heterogeneous compute, and new ways of connectivity, scale up and scale out connectivity.

Both compute stack as well as the infrastructure and connectivity stack is fundamentally different within less than 10 years, and it's moving very, very fast.

So because of all this transformation, brand new silicon architectures, brand new custom designs are being built for these types of infrastructure and systems.

And of course, AI is also making massive inroads into automotives, physical AI, into devices, edge.

So the entire stack of those devices is also being reinvented.

So I truly believe that we will see opportunities where the silicon stack, silicon architecture and software architecture is going to be pretty much redesigned for pretty much every segment in the industry.

The reason we are seeing more opportunities today in the data center, automotive infrastructure is That's where I think the heavy-duty, extremely high-performance compute is deployed first.