00:00 Speaker A
And Ferris was in the news this week. Announced that it has raised $50 million in a round led by Samsung Catalyst. What does that mean for your company, Ferris? How will you leverage that money?
00:11 Ferris
That’s a great question. So we call this an accelerator round. Um, and that primarily means expanding our commercial partnerships, um, growing our team. Well, like a lot of things in the past for us, it’s been about partnership. Well, it’s more deeply embedded in the strategic ecosystem than ever before. If you look at some of the investors that are participating, which includes not only Samsung, but also Micron Ventures, Celeste, founded by Intel’s current CEO Lee Boutin, and Eric Schmidt’s First Spark Ventures, I think you can kind of get an idea of the level of embedding we’re aiming for here.
00:54 Dan
Mr. Fair, Dan, here. What I would like to ask is, how do you differentiate your design process from what other companies have? Because you mentioned six. Obviously, Google has their own 6, and Broadcom is obviously there as well. There are different companies out there, we have Grock and their LPUs, and most likely from TSMC, but what sets your chip design apart from others?
01:22 Ferris
That’s a completely different design than what we’re aiming for. Thank you for your question, Dan. Oh, there are really two parts. One is that we are computing where the data resides, rather than having it shuttled back and forth between memory and the processor. This is an area that people call processing and memory, and it’s emerging as a very important area for solving energy problems and other kinds of efficiency problems. Second, we know that AI workloads such as video generation are already noisy and, in some sense, approximate. So our hardware deals with the noise rather than expending energy fighting it. Well, we’re not fighting physics, we’re leaning towards it.
02:11 Ferris
When you think about where traditional hardware is heading, efficiency is definitely increasing. But, you know, I think we’re still seeing things as predicted, even in the face of huge potential supply shortages, such as the rate at which hardware is advancing. Well, right now, if we keep going, we’re projected to need 134 gigawatts of electricity by 2030. So that’s a difference of about 50 million, sorry, 50 gigawatts. Um, what we’re demonstrating here is that there are alternative solutions to, um, finding creative ways to capture more energy and build nuclear power plants and things like that, and we can actually save orders of magnitude.
