David Gurra

And so you guys are basically saying throughput in terms of flops per millimeter squared, this is something you can own.

What was the breakthrough?

What is it that you're so good at to achieve this?

Yeah, so there's really a combination of two things.

If you look at the products in the market previously, there's been the HBM-based family, which is Nvidia, Google, Amazon, and then there's been the SRAM-based family.

And you are SRAM?

We are both, actually, uniquely.

So it's kind of taking two good ideas and putting them together.

It is possible to do both very high throughput, as you get from HBM, but also very low latency, as you get from SRAM, and do that in the same product.

What it gives you is actually a product that is better than any other product in the market at throughput.

This is exactly the flops per square millimeter that you described, while also matching some of the best, like the Cerebrus and the Grok at latency.

Let's talk about how quickly people could start deploying this right now, because what your aim is to complete the final design this year.

You hope to start manufacturing shipping even in 2027.

Who do you need to partner with on that?

How do you expect to be manufacturing here in the U.S.

or abroad?

Yeah, so I mean, there's a few big parts of the supply chain, and this is common for us as well as many other semiconductor firms in this space.

Really, you need logic wafers, memory wafers, which is HBM, and then you need rack build-outs.

And so those are the big parts of our supply chain.

TSMC is well-recognized as the best provider of logic wafers.